Membangun dan mengelola layanan tangguh menggunakan SDK AWS - Amazon Elastic Compute Cloud

Terjemahan disediakan oleh mesin penerjemah. Jika konten terjemahan yang diberikan bertentangan dengan versi bahasa Inggris aslinya, utamakan versi bahasa Inggris.

Membangun dan mengelola layanan tangguh menggunakan SDK AWS

Contoh kode berikut menunjukkan cara membuat layanan web dengan beban seimbang yang mengembalikan rekomendasi buku, film, dan lagu. Contoh ini menunjukkan cara layanan tersebut merespons kegagalan, serta cara merestrukturisasi layanan agar lebih tangguh ketika terjadi kegagalan.

  • Menggunakan grup Amazon EC2 Auto Scaling untuk membuat instans Amazon Elastic Compute Cloud (Amazon EC2) berdasarkan templat peluncuran dan menyimpan sejumlah instans dalam rentang yang ditentukan.

  • Menangani dan mendistribusikan permintaan HTTP dengan Elastic Load Balancing.

  • Memantau kondisi instans dalam grup Auto Scaling dan meneruskan permintaan hanya ke instans yang sehat.

  • Menjalankan server web Python pada setiap instans EC2 untuk menangani permintaan HTTP. Server web merespons dengan memberikan rekomendasi dan melakukan pemeriksaan kondisi.

  • Menyimulasikan layanan yang direkomendasikan dengan tabel Amazon DynamoDB.

  • Kontrol respons server web terhadap permintaan dan pemeriksaan kesehatan dengan memperbarui AWS Systems Manager parameter.

.NET
AWS SDK for .NET
catatan

Ada lebih banyak tentang GitHub. Temukan contoh lengkapnya dan pelajari cara pengaturan dan menjalankannya di Repositori Contoh Kode AWS.

Menjalankan skenario interaktif di prompt perintah.

static async Task Main(string[] args) { _configuration = new ConfigurationBuilder() .SetBasePath(Directory.GetCurrentDirectory()) .AddJsonFile("settings.json") // Load settings from .json file. .AddJsonFile("settings.local.json", true) // Optionally, load local settings. .Build(); // Set up dependency injection for the AWS services. using var host = Host.CreateDefaultBuilder(args) .ConfigureLogging(logging => logging.AddFilter("System", LogLevel.Debug) .AddFilter<DebugLoggerProvider>("Microsoft", LogLevel.Information) .AddFilter<ConsoleLoggerProvider>("Microsoft", LogLevel.Trace)) .ConfigureServices((_, services) => services.AddAWSService<IAmazonIdentityManagementService>() .AddAWSService<IAmazonDynamoDB>() .AddAWSService<IAmazonElasticLoadBalancingV2>() .AddAWSService<IAmazonSimpleSystemsManagement>() .AddAWSService<IAmazonAutoScaling>() .AddAWSService<IAmazonEC2>() .AddTransient<AutoScalerWrapper>() .AddTransient<ElasticLoadBalancerWrapper>() .AddTransient<SmParameterWrapper>() .AddTransient<Recommendations>() .AddSingleton<IConfiguration>(_configuration) ) .Build(); ServicesSetup(host); ResourcesSetup(); try { Console.WriteLine(new string('-', 80)); Console.WriteLine("Welcome to the Resilient Architecture Example Scenario."); Console.WriteLine(new string('-', 80)); await Deploy(true); Console.WriteLine("Now let's begin the scenario."); Console.WriteLine(new string('-', 80)); await Demo(true); Console.WriteLine(new string('-', 80)); Console.WriteLine("Finally, let's clean up our resources."); Console.WriteLine(new string('-', 80)); await DestroyResources(true); Console.WriteLine(new string('-', 80)); Console.WriteLine("Resilient Architecture Example Scenario is complete."); Console.WriteLine(new string('-', 80)); } catch (Exception ex) { Console.WriteLine(new string('-', 80)); Console.WriteLine($"There was a problem running the scenario: {ex.Message}"); await DestroyResources(true); Console.WriteLine(new string('-', 80)); } } /// <summary> /// Setup any common resources, also used for integration testing. /// </summary> public static void ResourcesSetup() { _httpClient = new HttpClient(); } /// <summary> /// Populate the services for use within the console application. /// </summary> /// <param name="host">The services host.</param> private static void ServicesSetup(IHost host) { _elasticLoadBalancerWrapper = host.Services.GetRequiredService<ElasticLoadBalancerWrapper>(); _iamClient = host.Services.GetRequiredService<IAmazonIdentityManagementService>(); _recommendations = host.Services.GetRequiredService<Recommendations>(); _autoScalerWrapper = host.Services.GetRequiredService<AutoScalerWrapper>(); _smParameterWrapper = host.Services.GetRequiredService<SmParameterWrapper>(); } /// <summary> /// Deploy necessary resources for the scenario. /// </summary> /// <param name="interactive">True to run as interactive.</param> /// <returns>True if successful.</returns> public static async Task<bool> Deploy(bool interactive) { var protocol = "HTTP"; var port = 80; var sshPort = 22; Console.WriteLine( "\nFor this demo, we'll use the AWS SDK for .NET to create several AWS resources\n" + "to set up a load-balanced web service endpoint and explore some ways to make it resilient\n" + "against various kinds of failures.\n\n" + "Some of the resources create by this demo are:\n"); Console.WriteLine( "\t* A DynamoDB table that the web service depends on to provide book, movie, and song recommendations."); Console.WriteLine( "\t* An EC2 launch template that defines EC2 instances that each contain a Python web server."); Console.WriteLine( "\t* An EC2 Auto Scaling group that manages EC2 instances across several Availability Zones."); Console.WriteLine( "\t* An Elastic Load Balancing (ELB) load balancer that targets the Auto Scaling group to distribute requests."); Console.WriteLine(new string('-', 80)); Console.WriteLine("Press Enter when you're ready to start deploying resources."); if (interactive) Console.ReadLine(); // Create and populate the DynamoDB table. var databaseTableName = _configuration["databaseName"]; var recommendationsPath = Path.Join(_configuration["resourcePath"], "recommendations_objects.json"); Console.WriteLine($"Creating and populating a DynamoDB table named {databaseTableName}."); await _recommendations.CreateDatabaseWithName(databaseTableName); await _recommendations.PopulateDatabase(databaseTableName, recommendationsPath); Console.WriteLine(new string('-', 80)); // Create the EC2 Launch Template. Console.WriteLine( $"Creating an EC2 launch template that runs 'server_startup_script.sh' when an instance starts.\n" + "\nThis script starts a Python web server defined in the `server.py` script. The web server\n" + "listens to HTTP requests on port 80 and responds to requests to '/' and to '/healthcheck'.\n" + "For demo purposes, this server is run as the root user. In production, the best practice is to\n" + "run a web server, such as Apache, with least-privileged credentials."); Console.WriteLine( "\nThe template also defines an IAM policy that each instance uses to assume a role that grants\n" + "permissions to access the DynamoDB recommendation table and Systems Manager parameters\n" + "that control the flow of the demo."); var startupScriptPath = Path.Join(_configuration["resourcePath"], "server_startup_script.sh"); var instancePolicyPath = Path.Join(_configuration["resourcePath"], "instance_policy.json"); await _autoScalerWrapper.CreateTemplate(startupScriptPath, instancePolicyPath); Console.WriteLine(new string('-', 80)); Console.WriteLine( "Creating an EC2 Auto Scaling group that maintains three EC2 instances, each in a different\n" + "Availability Zone.\n"); var zones = await _autoScalerWrapper.DescribeAvailabilityZones(); await _autoScalerWrapper.CreateGroupOfSize(3, _autoScalerWrapper.GroupName, zones); Console.WriteLine(new string('-', 80)); Console.WriteLine( "At this point, you have EC2 instances created. Once each instance starts, it listens for\n" + "HTTP requests. You can see these instances in the console or continue with the demo.\n"); Console.WriteLine(new string('-', 80)); Console.WriteLine("Press Enter when you're ready to continue."); if (interactive) Console.ReadLine(); Console.WriteLine("Creating variables that control the flow of the demo."); await _smParameterWrapper.Reset(); Console.WriteLine( "\nCreating an Elastic Load Balancing target group and load balancer. The target group\n" + "defines how the load balancer connects to instances. The load balancer provides a\n" + "single endpoint where clients connect and dispatches requests to instances in the group."); var defaultVpc = await _autoScalerWrapper.GetDefaultVpc(); var subnets = await _autoScalerWrapper.GetAllVpcSubnetsForZones(defaultVpc.VpcId, zones); var subnetIds = subnets.Select(s => s.SubnetId).ToList(); var targetGroup = await _elasticLoadBalancerWrapper.CreateTargetGroupOnVpc(_elasticLoadBalancerWrapper.TargetGroupName, protocol, port, defaultVpc.VpcId); await _elasticLoadBalancerWrapper.CreateLoadBalancerAndListener(_elasticLoadBalancerWrapper.LoadBalancerName, subnetIds, targetGroup); await _autoScalerWrapper.AttachLoadBalancerToGroup(_autoScalerWrapper.GroupName, targetGroup.TargetGroupArn); Console.WriteLine("\nVerifying access to the load balancer endpoint..."); var endPoint = await _elasticLoadBalancerWrapper.GetEndpointForLoadBalancerByName(_elasticLoadBalancerWrapper.LoadBalancerName); var loadBalancerAccess = await _elasticLoadBalancerWrapper.VerifyLoadBalancerEndpoint(endPoint); if (!loadBalancerAccess) { Console.WriteLine("\nCouldn't connect to the load balancer, verifying that the port is open..."); var ipString = await _httpClient.GetStringAsync("https://checkip.amazonaws.com"); ipString = ipString.Trim(); var defaultSecurityGroup = await _autoScalerWrapper.GetDefaultSecurityGroupForVpc(defaultVpc); var portIsOpen = _autoScalerWrapper.VerifyInboundPortForGroup(defaultSecurityGroup, port, ipString); var sshPortIsOpen = _autoScalerWrapper.VerifyInboundPortForGroup(defaultSecurityGroup, sshPort, ipString); if (!portIsOpen) { Console.WriteLine( "\nFor this example to work, the default security group for your default VPC must\n" + "allows access from this computer. You can either add it automatically from this\n" + "example or add it yourself using the AWS Management Console.\n"); if (!interactive || GetYesNoResponse( "Do you want to add a rule to the security group to allow inbound traffic from your computer's IP address?")) { await _autoScalerWrapper.OpenInboundPort(defaultSecurityGroup.GroupId, port, ipString); } } if (!sshPortIsOpen) { if (!interactive || GetYesNoResponse( "Do you want to add a rule to the security group to allow inbound SSH traffic for debugging from your computer's IP address?")) { await _autoScalerWrapper.OpenInboundPort(defaultSecurityGroup.GroupId, sshPort, ipString); } } loadBalancerAccess = await _elasticLoadBalancerWrapper.VerifyLoadBalancerEndpoint(endPoint); } if (loadBalancerAccess) { Console.WriteLine("Your load balancer is ready. You can access it by browsing to:"); Console.WriteLine($"\thttp://{endPoint}\n"); } else { Console.WriteLine( "\nCouldn't get a successful response from the load balancer endpoint. Troubleshoot by\n" + "manually verifying that your VPC and security group are configured correctly and that\n" + "you can successfully make a GET request to the load balancer endpoint:\n"); Console.WriteLine($"\thttp://{endPoint}\n"); } Console.WriteLine(new string('-', 80)); Console.WriteLine("Press Enter when you're ready to continue with the demo."); if (interactive) Console.ReadLine(); return true; } /// <summary> /// Demonstrate the steps of the scenario. /// </summary> /// <param name="interactive">True to run as an interactive scenario.</param> /// <returns>Async task.</returns> public static async Task<bool> Demo(bool interactive) { var ssmOnlyPolicy = Path.Join(_configuration["resourcePath"], "ssm_only_policy.json"); Console.WriteLine(new string('-', 80)); Console.WriteLine("Resetting parameters to starting values for demo."); await _smParameterWrapper.Reset(); Console.WriteLine("\nThis part of the demonstration shows how to toggle different parts of the system\n" + "to create situations where the web service fails, and shows how using a resilient\n" + "architecture can keep the web service running in spite of these failures."); Console.WriteLine(new string('-', 88)); Console.WriteLine("At the start, the load balancer endpoint returns recommendations and reports that all targets are healthy."); if (interactive) await DemoActionChoices(); Console.WriteLine($"The web service running on the EC2 instances gets recommendations by querying a DynamoDB table.\n" + $"The table name is contained in a Systems Manager parameter named '{_smParameterWrapper.TableParameter}'.\n" + $"To simulate a failure of the recommendation service, let's set this parameter to name a non-existent table.\n"); await _smParameterWrapper.PutParameterByName(_smParameterWrapper.TableParameter, "this-is-not-a-table"); Console.WriteLine("\nNow, sending a GET request to the load balancer endpoint returns a failure code. But, the service reports as\n" + "healthy to the load balancer because shallow health checks don't check for failure of the recommendation service."); if (interactive) await DemoActionChoices(); Console.WriteLine("Instead of failing when the recommendation service fails, the web service can return a static response."); Console.WriteLine("While this is not a perfect solution, it presents the customer with a somewhat better experience than failure."); await _smParameterWrapper.PutParameterByName(_smParameterWrapper.FailureResponseParameter, "static"); Console.WriteLine("\nNow, sending a GET request to the load balancer endpoint returns a static response."); Console.WriteLine("The service still reports as healthy because health checks are still shallow."); if (interactive) await DemoActionChoices(); Console.WriteLine("Let's reinstate the recommendation service.\n"); await _smParameterWrapper.PutParameterByName(_smParameterWrapper.TableParameter, _smParameterWrapper.TableName); Console.WriteLine( "\nLet's also substitute bad credentials for one of the instances in the target group so that it can't\n" + "access the DynamoDB recommendation table.\n" ); await _autoScalerWrapper.CreateInstanceProfileWithName( _autoScalerWrapper.BadCredsPolicyName, _autoScalerWrapper.BadCredsRoleName, _autoScalerWrapper.BadCredsProfileName, ssmOnlyPolicy, new List<string> { "AmazonSSMManagedInstanceCore" } ); var instances = await _autoScalerWrapper.GetInstancesByGroupName(_autoScalerWrapper.GroupName); var badInstanceId = instances.First(); var instanceProfile = await _autoScalerWrapper.GetInstanceProfile(badInstanceId); Console.WriteLine( $"Replacing the profile for instance {badInstanceId} with a profile that contains\n" + "bad credentials...\n" ); await _autoScalerWrapper.ReplaceInstanceProfile( badInstanceId, _autoScalerWrapper.BadCredsProfileName, instanceProfile.AssociationId ); Console.WriteLine( "Now, sending a GET request to the load balancer endpoint returns either a recommendation or a static response,\n" + "depending on which instance is selected by the load balancer.\n" ); if (interactive) await DemoActionChoices(); Console.WriteLine("\nLet's implement a deep health check. For this demo, a deep health check tests whether"); Console.WriteLine("the web service can access the DynamoDB table that it depends on for recommendations. Note that"); Console.WriteLine("the deep health check is only for ELB routing and not for Auto Scaling instance health."); Console.WriteLine("This kind of deep health check is not recommended for Auto Scaling instance health, because it"); Console.WriteLine("risks accidental termination of all instances in the Auto Scaling group when a dependent service fails."); Console.WriteLine("\nBy implementing deep health checks, the load balancer can detect when one of the instances is failing"); Console.WriteLine("and take that instance out of rotation."); await _smParameterWrapper.PutParameterByName(_smParameterWrapper.HealthCheckParameter, "deep"); Console.WriteLine($"\nNow, checking target health indicates that the instance with bad credentials ({badInstanceId})"); Console.WriteLine("is unhealthy. Note that it might take a minute or two for the load balancer to detect the unhealthy"); Console.WriteLine("instance. Sending a GET request to the load balancer endpoint always returns a recommendation, because"); Console.WriteLine("the load balancer takes unhealthy instances out of its rotation."); if (interactive) await DemoActionChoices(); Console.WriteLine("\nBecause the instances in this demo are controlled by an auto scaler, the simplest way to fix an unhealthy"); Console.WriteLine("instance is to terminate it and let the auto scaler start a new instance to replace it."); await _autoScalerWrapper.TryTerminateInstanceById(badInstanceId); Console.WriteLine($"\nEven while the instance is terminating and the new instance is starting, sending a GET"); Console.WriteLine("request to the web service continues to get a successful recommendation response because"); Console.WriteLine("starts and reports as healthy, it is included in the load balancing rotation."); Console.WriteLine("Note that terminating and replacing an instance typically takes several minutes, during which time you"); Console.WriteLine("can see the changing health check status until the new instance is running and healthy."); if (interactive) await DemoActionChoices(); Console.WriteLine("\nIf the recommendation service fails now, deep health checks mean all instances report as unhealthy."); await _smParameterWrapper.PutParameterByName(_smParameterWrapper.TableParameter, "this-is-not-a-table"); Console.WriteLine($"\nWhen all instances are unhealthy, the load balancer continues to route requests even to"); Console.WriteLine("unhealthy instances, allowing them to fail open and return a static response rather than fail"); Console.WriteLine("closed and report failure to the customer."); if (interactive) await DemoActionChoices(); await _smParameterWrapper.Reset(); Console.WriteLine(new string('-', 80)); return true; } /// <summary> /// Clean up the resources from the scenario. /// </summary> /// <param name="interactive">True to ask the user for cleanup.</param> /// <returns>Async task.</returns> public static async Task<bool> DestroyResources(bool interactive) { Console.WriteLine(new string('-', 80)); Console.WriteLine( "To keep things tidy and to avoid unwanted charges on your account, we can clean up all AWS resources\n" + "that were created for this demo." ); if (!interactive || GetYesNoResponse("Do you want to clean up all demo resources? (y/n) ")) { await _elasticLoadBalancerWrapper.DeleteLoadBalancerByName(_elasticLoadBalancerWrapper.LoadBalancerName); await _elasticLoadBalancerWrapper.DeleteTargetGroupByName(_elasticLoadBalancerWrapper.TargetGroupName); await _autoScalerWrapper.TerminateAndDeleteAutoScalingGroupWithName(_autoScalerWrapper.GroupName); await _autoScalerWrapper.DeleteKeyPairByName(_autoScalerWrapper.KeyPairName); await _autoScalerWrapper.DeleteTemplateByName(_autoScalerWrapper.LaunchTemplateName); await _autoScalerWrapper.DeleteInstanceProfile( _autoScalerWrapper.BadCredsProfileName, _autoScalerWrapper.BadCredsRoleName ); await _recommendations.DestroyDatabaseByName(_recommendations.TableName); } else { Console.WriteLine( "Ok, we'll leave the resources intact.\n" + "Don't forget to delete them when you're done with them or you might incur unexpected charges." ); } Console.WriteLine(new string('-', 80)); return true; }

Membuat kelas yang menggabungkan tindakan Penskalaan Otomatis dan Amazon EC2.

/// <summary> /// Encapsulates Amazon EC2 Auto Scaling and EC2 management methods. /// </summary> public class AutoScalerWrapper { private readonly IAmazonAutoScaling _amazonAutoScaling; private readonly IAmazonEC2 _amazonEc2; private readonly IAmazonSimpleSystemsManagement _amazonSsm; private readonly IAmazonIdentityManagementService _amazonIam; private readonly string _instanceType = ""; private readonly string _amiParam = ""; private readonly string _launchTemplateName = ""; private readonly string _groupName = ""; private readonly string _instancePolicyName = ""; private readonly string _instanceRoleName = ""; private readonly string _instanceProfileName = ""; private readonly string _badCredsProfileName = ""; private readonly string _badCredsRoleName = ""; private readonly string _badCredsPolicyName = ""; private readonly string _keyPairName = ""; public string GroupName => _groupName; public string KeyPairName => _keyPairName; public string LaunchTemplateName => _launchTemplateName; public string InstancePolicyName => _instancePolicyName; public string BadCredsProfileName => _badCredsProfileName; public string BadCredsRoleName => _badCredsRoleName; public string BadCredsPolicyName => _badCredsPolicyName; /// <summary> /// Constructor for the AutoScalerWrapper. /// </summary> /// <param name="amazonAutoScaling">The injected AutoScaling client.</param> /// <param name="amazonEc2">The injected EC2 client.</param> /// <param name="amazonIam">The injected IAM client.</param> /// <param name="amazonSsm">The injected SSM client.</param> public AutoScalerWrapper( IAmazonAutoScaling amazonAutoScaling, IAmazonEC2 amazonEc2, IAmazonSimpleSystemsManagement amazonSsm, IAmazonIdentityManagementService amazonIam, IConfiguration configuration) { _amazonAutoScaling = amazonAutoScaling; _amazonEc2 = amazonEc2; _amazonSsm = amazonSsm; _amazonIam = amazonIam; var prefix = configuration["resourcePrefix"]; _instanceType = configuration["instanceType"]; _amiParam = configuration["amiParam"]; _launchTemplateName = prefix + "-template"; _groupName = prefix + "-group"; _instancePolicyName = prefix + "-pol"; _instanceRoleName = prefix + "-role"; _instanceProfileName = prefix + "-prof"; _badCredsPolicyName = prefix + "-bc-pol"; _badCredsRoleName = prefix + "-bc-role"; _badCredsProfileName = prefix + "-bc-prof"; _keyPairName = prefix + "-key-pair"; } /// <summary> /// Create a policy, role, and profile that is associated with instances with a specified name. /// An instance's associated profile defines a role that is assumed by the /// instance.The role has attached policies that specify the AWS permissions granted to /// clients that run on the instance. /// </summary> /// <param name="policyName">Name to use for the policy.</param> /// <param name="roleName">Name to use for the role.</param> /// <param name="profileName">Name to use for the profile.</param> /// <param name="ssmOnlyPolicyFile">Path to a policy file for SSM.</param> /// <param name="awsManagedPolicies">AWS Managed policies to be attached to the role.</param> /// <returns>The Arn of the profile.</returns> public async Task<string> CreateInstanceProfileWithName( string policyName, string roleName, string profileName, string ssmOnlyPolicyFile, List<string>? awsManagedPolicies = null) { var assumeRoleDoc = "{" + "\"Version\": \"2012-10-17\"," + "\"Statement\": [{" + "\"Effect\": \"Allow\"," + "\"Principal\": {" + "\"Service\": [" + "\"ec2.amazonaws.com\"" + "]" + "}," + "\"Action\": \"sts:AssumeRole\"" + "}]" + "}"; var policyDocument = await File.ReadAllTextAsync(ssmOnlyPolicyFile); var policyArn = ""; try { var createPolicyResult = await _amazonIam.CreatePolicyAsync( new CreatePolicyRequest { PolicyName = policyName, PolicyDocument = policyDocument }); policyArn = createPolicyResult.Policy.Arn; } catch (EntityAlreadyExistsException) { // The policy already exists, so we look it up to get the Arn. var policiesPaginator = _amazonIam.Paginators.ListPolicies( new ListPoliciesRequest() { Scope = PolicyScopeType.Local }); // Get the entire list using the paginator. await foreach (var policy in policiesPaginator.Policies) { if (policy.PolicyName.Equals(policyName)) { policyArn = policy.Arn; } } if (policyArn == null) { throw new InvalidOperationException("Policy not found"); } } try { await _amazonIam.CreateRoleAsync(new CreateRoleRequest() { RoleName = roleName, AssumeRolePolicyDocument = assumeRoleDoc, }); await _amazonIam.AttachRolePolicyAsync(new AttachRolePolicyRequest() { RoleName = roleName, PolicyArn = policyArn }); if (awsManagedPolicies != null) { foreach (var awsPolicy in awsManagedPolicies) { await _amazonIam.AttachRolePolicyAsync(new AttachRolePolicyRequest() { PolicyArn = $"arn:aws:iam::aws:policy/{awsPolicy}", RoleName = roleName }); } } } catch (EntityAlreadyExistsException) { Console.WriteLine("Role already exists."); } string profileArn = ""; try { var profileCreateResponse = await _amazonIam.CreateInstanceProfileAsync( new CreateInstanceProfileRequest() { InstanceProfileName = profileName }); // Allow time for the profile to be ready. profileArn = profileCreateResponse.InstanceProfile.Arn; Thread.Sleep(10000); await _amazonIam.AddRoleToInstanceProfileAsync( new AddRoleToInstanceProfileRequest() { InstanceProfileName = profileName, RoleName = roleName }); } catch (EntityAlreadyExistsException) { Console.WriteLine("Policy already exists."); var profileGetResponse = await _amazonIam.GetInstanceProfileAsync( new GetInstanceProfileRequest() { InstanceProfileName = profileName }); profileArn = profileGetResponse.InstanceProfile.Arn; } return profileArn; } /// <summary> /// Create a new key pair and save the file. /// </summary> /// <param name="newKeyPairName">The name of the new key pair.</param> /// <returns>Async task.</returns> public async Task CreateKeyPair(string newKeyPairName) { try { var keyResponse = await _amazonEc2.CreateKeyPairAsync( new CreateKeyPairRequest() { KeyName = newKeyPairName }); await File.WriteAllTextAsync($"{newKeyPairName}.pem", keyResponse.KeyPair.KeyMaterial); Console.WriteLine($"Created key pair {newKeyPairName}."); } catch (AlreadyExistsException) { Console.WriteLine("Key pair already exists."); } } /// <summary> /// Delete the key pair and file by name. /// </summary> /// <param name="deleteKeyPairName">The key pair to delete.</param> /// <returns>Async task.</returns> public async Task DeleteKeyPairByName(string deleteKeyPairName) { try { await _amazonEc2.DeleteKeyPairAsync( new DeleteKeyPairRequest() { KeyName = deleteKeyPairName }); File.Delete($"{deleteKeyPairName}.pem"); } catch (FileNotFoundException) { Console.WriteLine($"Key pair {deleteKeyPairName} not found."); } } /// <summary> /// Creates an Amazon EC2 launch template to use with Amazon EC2 Auto Scaling. /// The launch template specifies a Bash script in its user data field that runs after /// the instance is started. This script installs the Python packages and starts a Python /// web server on the instance. /// </summary> /// <param name="startupScriptPath">The path to a Bash script file that is run.</param> /// <param name="instancePolicyPath">The path to a permissions policy to create and attach to the profile.</param> /// <returns>The template object.</returns> public async Task<Amazon.EC2.Model.LaunchTemplate> CreateTemplate(string startupScriptPath, string instancePolicyPath) { await CreateKeyPair(_keyPairName); await CreateInstanceProfileWithName(_instancePolicyName, _instanceRoleName, _instanceProfileName, instancePolicyPath); var startServerText = await File.ReadAllTextAsync(startupScriptPath); var plainTextBytes = System.Text.Encoding.UTF8.GetBytes(startServerText); var amiLatest = await _amazonSsm.GetParameterAsync( new GetParameterRequest() { Name = _amiParam }); var amiId = amiLatest.Parameter.Value; var launchTemplateResponse = await _amazonEc2.CreateLaunchTemplateAsync( new CreateLaunchTemplateRequest() { LaunchTemplateName = _launchTemplateName, LaunchTemplateData = new RequestLaunchTemplateData() { InstanceType = _instanceType, ImageId = amiId, IamInstanceProfile = new LaunchTemplateIamInstanceProfileSpecificationRequest() { Name = _instanceProfileName }, KeyName = _keyPairName, UserData = System.Convert.ToBase64String(plainTextBytes) } }); return launchTemplateResponse.LaunchTemplate; } /// <summary> /// Get a list of Availability Zones in the AWS Region of the Amazon EC2 Client. /// </summary> /// <returns>A list of availability zones.</returns> public async Task<List<string>> DescribeAvailabilityZones() { var zoneResponse = await _amazonEc2.DescribeAvailabilityZonesAsync( new DescribeAvailabilityZonesRequest()); return zoneResponse.AvailabilityZones.Select(z => z.ZoneName).ToList(); } /// <summary> /// Create an EC2 Auto Scaling group of a specified size and name. /// </summary> /// <param name="groupSize">The size for the group.</param> /// <param name="groupName">The name for the group.</param> /// <param name="availabilityZones">The availability zones for the group.</param> /// <returns>Async task.</returns> public async Task CreateGroupOfSize(int groupSize, string groupName, List<string> availabilityZones) { try { await _amazonAutoScaling.CreateAutoScalingGroupAsync( new CreateAutoScalingGroupRequest() { AutoScalingGroupName = groupName, AvailabilityZones = availabilityZones, LaunchTemplate = new Amazon.AutoScaling.Model.LaunchTemplateSpecification() { LaunchTemplateName = _launchTemplateName, Version = "$Default" }, MaxSize = groupSize, MinSize = groupSize }); Console.WriteLine($"Created EC2 Auto Scaling group {groupName} with size {groupSize}."); } catch (EntityAlreadyExistsException) { Console.WriteLine($"EC2 Auto Scaling group {groupName} already exists."); } } /// <summary> /// Get the default VPC for the account. /// </summary> /// <returns>The default VPC object.</returns> public async Task<Vpc> GetDefaultVpc() { var vpcResponse = await _amazonEc2.DescribeVpcsAsync( new DescribeVpcsRequest() { Filters = new List<Amazon.EC2.Model.Filter>() { new ("is-default", new List<string>() { "true" }) } }); return vpcResponse.Vpcs[0]; } /// <summary> /// Get all the subnets for a Vpc in a set of availability zones. /// </summary> /// <param name="vpcId">The Id of the Vpc.</param> /// <param name="availabilityZones">The list of availability zones.</param> /// <returns>The collection of subnet objects.</returns> public async Task<List<Subnet>> GetAllVpcSubnetsForZones(string vpcId, List<string> availabilityZones) { var subnets = new List<Subnet>(); var subnetPaginator = _amazonEc2.Paginators.DescribeSubnets( new DescribeSubnetsRequest() { Filters = new List<Amazon.EC2.Model.Filter>() { new ("vpc-id", new List<string>() { vpcId}), new ("availability-zone", availabilityZones), new ("default-for-az", new List<string>() { "true" }) } }); // Get the entire list using the paginator. await foreach (var subnet in subnetPaginator.Subnets) { subnets.Add(subnet); } return subnets; } /// <summary> /// Delete a launch template by name. /// </summary> /// <param name="templateName">The name of the template to delete.</param> /// <returns>Async task.</returns> public async Task DeleteTemplateByName(string templateName) { try { await _amazonEc2.DeleteLaunchTemplateAsync( new DeleteLaunchTemplateRequest() { LaunchTemplateName = templateName }); } catch (AmazonClientException) { Console.WriteLine($"Unable to delete template {templateName}."); } } /// <summary> /// Detaches a role from an instance profile, detaches policies from the role, /// and deletes all the resources. /// </summary> /// <param name="profileName">The name of the profile to delete.</param> /// <param name="roleName">The name of the role to delete.</param> /// <returns>Async task.</returns> public async Task DeleteInstanceProfile(string profileName, string roleName) { try { await _amazonIam.RemoveRoleFromInstanceProfileAsync( new RemoveRoleFromInstanceProfileRequest() { InstanceProfileName = profileName, RoleName = roleName }); await _amazonIam.DeleteInstanceProfileAsync( new DeleteInstanceProfileRequest() { InstanceProfileName = profileName }); var attachedPolicies = await _amazonIam.ListAttachedRolePoliciesAsync( new ListAttachedRolePoliciesRequest() { RoleName = roleName }); foreach (var policy in attachedPolicies.AttachedPolicies) { await _amazonIam.DetachRolePolicyAsync( new DetachRolePolicyRequest() { RoleName = roleName, PolicyArn = policy.PolicyArn }); // Delete the custom policies only. if (!policy.PolicyArn.StartsWith("arn:aws:iam::aws")) { await _amazonIam.DeletePolicyAsync( new Amazon.IdentityManagement.Model.DeletePolicyRequest() { PolicyArn = policy.PolicyArn }); } } await _amazonIam.DeleteRoleAsync( new DeleteRoleRequest() { RoleName = roleName }); } catch (NoSuchEntityException) { Console.WriteLine($"Instance profile {profileName} does not exist."); } } /// <summary> /// Gets data about the instances in an EC2 Auto Scaling group by its group name. /// </summary> /// <param name="group">The name of the auto scaling group.</param> /// <returns>A collection of instance Ids.</returns> public async Task<IEnumerable<string>> GetInstancesByGroupName(string group) { var instanceResponse = await _amazonAutoScaling.DescribeAutoScalingGroupsAsync( new DescribeAutoScalingGroupsRequest() { AutoScalingGroupNames = new List<string>() { group } }); var instanceIds = instanceResponse.AutoScalingGroups.SelectMany( g => g.Instances.Select(i => i.InstanceId)); return instanceIds; } /// <summary> /// Get the instance profile association data for an instance. /// </summary> /// <param name="instanceId">The Id of the instance.</param> /// <returns>Instance profile associations data.</returns> public async Task<IamInstanceProfileAssociation> GetInstanceProfile(string instanceId) { var response = await _amazonEc2.DescribeIamInstanceProfileAssociationsAsync( new DescribeIamInstanceProfileAssociationsRequest() { Filters = new List<Amazon.EC2.Model.Filter>() { new ("instance-id", new List<string>() { instanceId }) }, }); return response.IamInstanceProfileAssociations[0]; } /// <summary> /// Replace the profile associated with a running instance. After the profile is replaced, the instance /// is rebooted to ensure that it uses the new profile. When the instance is ready, Systems Manager is /// used to restart the Python web server. /// </summary> /// <param name="instanceId">The Id of the instance to update.</param> /// <param name="credsProfileName">The name of the new profile to associate with the specified instance.</param> /// <param name="associationId">The Id of the existing profile association for the instance.</param> /// <returns>Async task.</returns> public async Task ReplaceInstanceProfile(string instanceId, string credsProfileName, string associationId) { await _amazonEc2.ReplaceIamInstanceProfileAssociationAsync( new ReplaceIamInstanceProfileAssociationRequest() { AssociationId = associationId, IamInstanceProfile = new IamInstanceProfileSpecification() { Name = credsProfileName } }); // Allow time before resetting. Thread.Sleep(25000); var instanceReady = false; var retries = 5; while (retries-- > 0 && !instanceReady) { await _amazonEc2.RebootInstancesAsync( new RebootInstancesRequest(new List<string>() { instanceId })); Thread.Sleep(10000); var instancesPaginator = _amazonSsm.Paginators.DescribeInstanceInformation( new DescribeInstanceInformationRequest()); // Get the entire list using the paginator. await foreach (var instance in instancesPaginator.InstanceInformationList) { instanceReady = instance.InstanceId == instanceId; if (instanceReady) { break; } } } Console.WriteLine($"Sending restart command to instance {instanceId}"); await _amazonSsm.SendCommandAsync( new SendCommandRequest() { InstanceIds = new List<string>() { instanceId }, DocumentName = "AWS-RunShellScript", Parameters = new Dictionary<string, List<string>>() { {"commands", new List<string>() { "cd / && sudo python3 server.py 80" }} } }); Console.WriteLine($"Restarted the web server on instance {instanceId}"); } /// <summary> /// Try to terminate an instance by its Id. /// </summary> /// <param name="instanceId">The Id of the instance to terminate.</param> /// <returns>Async task.</returns> public async Task TryTerminateInstanceById(string instanceId) { var stopping = false; Console.WriteLine($"Stopping {instanceId}..."); while (!stopping) { try { await _amazonAutoScaling.TerminateInstanceInAutoScalingGroupAsync( new TerminateInstanceInAutoScalingGroupRequest() { InstanceId = instanceId, ShouldDecrementDesiredCapacity = false }); stopping = true; } catch (ScalingActivityInProgressException) { Console.WriteLine($"Scaling activity in progress for {instanceId}. Waiting..."); Thread.Sleep(10000); } } } /// <summary> /// Tries to delete the EC2 Auto Scaling group. If the group is in use or in progress, /// waits and retries until the group is successfully deleted. /// </summary> /// <param name="groupName">The name of the group to try to delete.</param> /// <returns>Async task.</returns> public async Task TryDeleteGroupByName(string groupName) { var stopped = false; while (!stopped) { try { await _amazonAutoScaling.DeleteAutoScalingGroupAsync( new DeleteAutoScalingGroupRequest() { AutoScalingGroupName = groupName }); stopped = true; } catch (Exception e) when ((e is ScalingActivityInProgressException) || (e is Amazon.AutoScaling.Model.ResourceInUseException)) { Console.WriteLine($"Some instances are still running. Waiting..."); Thread.Sleep(10000); } } } /// <summary> /// Terminate instances and delete the Auto Scaling group by name. /// </summary> /// <param name="groupName">The name of the group to delete.</param> /// <returns>Async task.</returns> public async Task TerminateAndDeleteAutoScalingGroupWithName(string groupName) { var describeGroupsResponse = await _amazonAutoScaling.DescribeAutoScalingGroupsAsync( new DescribeAutoScalingGroupsRequest() { AutoScalingGroupNames = new List<string>() { groupName } }); if (describeGroupsResponse.AutoScalingGroups.Any()) { // Update the size to 0. await _amazonAutoScaling.UpdateAutoScalingGroupAsync( new UpdateAutoScalingGroupRequest() { AutoScalingGroupName = groupName, MinSize = 0 }); var group = describeGroupsResponse.AutoScalingGroups[0]; foreach (var instance in group.Instances) { await TryTerminateInstanceById(instance.InstanceId); } await TryDeleteGroupByName(groupName); } else { Console.WriteLine($"No groups found with name {groupName}."); } } /// <summary> /// Get the default security group for a specified Vpc. /// </summary> /// <param name="vpc">The Vpc to search.</param> /// <returns>The default security group.</returns> public async Task<SecurityGroup> GetDefaultSecurityGroupForVpc(Vpc vpc) { var groupResponse = await _amazonEc2.DescribeSecurityGroupsAsync( new DescribeSecurityGroupsRequest() { Filters = new List<Amazon.EC2.Model.Filter>() { new ("group-name", new List<string>() { "default" }), new ("vpc-id", new List<string>() { vpc.VpcId }) } }); return groupResponse.SecurityGroups[0]; } /// <summary> /// Verify the default security group of a Vpc allows ingress from the calling computer. /// This can be done by allowing ingress from this computer's IP address. /// In some situations, such as connecting from a corporate network, you must instead specify /// a prefix list Id. You can also temporarily open the port to any IP address while running this example. /// If you do, be sure to remove public access when you're done. /// </summary> /// <param name="vpc">The group to check.</param> /// <param name="port">The port to verify.</param> /// <param name="ipAddress">This computer's IP address.</param> /// <returns>True if the ip address is allowed on the group.</returns> public bool VerifyInboundPortForGroup(SecurityGroup group, int port, string ipAddress) { var portIsOpen = false; foreach (var ipPermission in group.IpPermissions) { if (ipPermission.FromPort == port) { foreach (var ipRange in ipPermission.Ipv4Ranges) { var cidr = ipRange.CidrIp; if (cidr.StartsWith(ipAddress) || cidr == "0.0.0.0/0") { portIsOpen = true; } } if (ipPermission.PrefixListIds.Any()) { portIsOpen = true; } if (!portIsOpen) { Console.WriteLine("The inbound rule does not appear to be open to either this computer's IP\n" + "address, to all IP addresses (0.0.0.0/0), or to a prefix list ID."); } else { break; } } } return portIsOpen; } /// <summary> /// Add an ingress rule to the specified security group that allows access on the /// specified port from the specified IP address. /// </summary> /// <param name="groupId">The Id of the security group to modify.</param> /// <param name="port">The port to open.</param> /// <param name="ipAddress">The IP address to allow access.</param> /// <returns>Async task.</returns> public async Task OpenInboundPort(string groupId, int port, string ipAddress) { await _amazonEc2.AuthorizeSecurityGroupIngressAsync( new AuthorizeSecurityGroupIngressRequest() { GroupId = groupId, IpPermissions = new List<IpPermission>() { new IpPermission() { FromPort = port, ToPort = port, IpProtocol = "tcp", Ipv4Ranges = new List<IpRange>() { new IpRange() { CidrIp = $"{ipAddress}/32" } } } } }); } /// <summary> /// Attaches an Elastic Load Balancing (ELB) target group to this EC2 Auto Scaling group. /// The /// </summary> /// <param name="autoScalingGroupName">The name of the Auto Scaling group.</param> /// <param name="targetGroupArn">The Arn for the target group.</param> /// <returns>Async task.</returns> public async Task AttachLoadBalancerToGroup(string autoScalingGroupName, string targetGroupArn) { await _amazonAutoScaling.AttachLoadBalancerTargetGroupsAsync( new AttachLoadBalancerTargetGroupsRequest() { AutoScalingGroupName = autoScalingGroupName, TargetGroupARNs = new List<string>() { targetGroupArn } }); } }

Membuat kelas yang menggabungkan tindakan Penyeimbangan Beban Elastis.

/// <summary> /// Encapsulates Elastic Load Balancer actions. /// </summary> public class ElasticLoadBalancerWrapper { private readonly IAmazonElasticLoadBalancingV2 _amazonElasticLoadBalancingV2; private string? _endpoint = null; private readonly string _targetGroupName = ""; private readonly string _loadBalancerName = ""; HttpClient _httpClient = new(); public string TargetGroupName => _targetGroupName; public string LoadBalancerName => _loadBalancerName; /// <summary> /// Constructor for the Elastic Load Balancer wrapper. /// </summary> /// <param name="amazonElasticLoadBalancingV2">The injected load balancing v2 client.</param> /// <param name="configuration">The injected configuration.</param> public ElasticLoadBalancerWrapper( IAmazonElasticLoadBalancingV2 amazonElasticLoadBalancingV2, IConfiguration configuration) { _amazonElasticLoadBalancingV2 = amazonElasticLoadBalancingV2; var prefix = configuration["resourcePrefix"]; _targetGroupName = prefix + "-tg"; _loadBalancerName = prefix + "-lb"; } /// <summary> /// Get the HTTP Endpoint of a load balancer by its name. /// </summary> /// <param name="loadBalancerName">The name of the load balancer.</param> /// <returns>The HTTP endpoint.</returns> public async Task<string> GetEndpointForLoadBalancerByName(string loadBalancerName) { if (_endpoint == null) { var endpointResponse = await _amazonElasticLoadBalancingV2.DescribeLoadBalancersAsync( new DescribeLoadBalancersRequest() { Names = new List<string>() { loadBalancerName } }); _endpoint = endpointResponse.LoadBalancers[0].DNSName; } return _endpoint; } /// <summary> /// Return the GET response for an endpoint as text. /// </summary> /// <param name="endpoint">The endpoint for the request.</param> /// <returns>The request response.</returns> public async Task<string> GetEndPointResponse(string endpoint) { var endpointResponse = await _httpClient.GetAsync($"http://{endpoint}"); var textResponse = await endpointResponse.Content.ReadAsStringAsync(); return textResponse!; } /// <summary> /// Get the target health for a group by name. /// </summary> /// <param name="groupName">The name of the group.</param> /// <returns>The collection of health descriptions.</returns> public async Task<List<TargetHealthDescription>> CheckTargetHealthForGroup(string groupName) { List<TargetHealthDescription> result = null!; try { var groupResponse = await _amazonElasticLoadBalancingV2.DescribeTargetGroupsAsync( new DescribeTargetGroupsRequest() { Names = new List<string>() { groupName } }); var healthResponse = await _amazonElasticLoadBalancingV2.DescribeTargetHealthAsync( new DescribeTargetHealthRequest() { TargetGroupArn = groupResponse.TargetGroups[0].TargetGroupArn }); ; result = healthResponse.TargetHealthDescriptions; } catch (TargetGroupNotFoundException) { Console.WriteLine($"Target group {groupName} not found."); } return result; } /// <summary> /// Create an Elastic Load Balancing target group. The target group specifies how the load balancer forwards /// requests to instances in the group and how instance health is checked. /// /// To speed up this demo, the health check is configured with shortened times and lower thresholds. In production, /// you might want to decrease the sensitivity of your health checks to avoid unwanted failures. /// </summary> /// <param name="groupName">The name for the group.</param> /// <param name="protocol">The protocol, such as HTTP.</param> /// <param name="port">The port to use to forward requests, such as 80.</param> /// <param name="vpcId">The Id of the Vpc in which the load balancer exists.</param> /// <returns>The new TargetGroup object.</returns> public async Task<TargetGroup> CreateTargetGroupOnVpc(string groupName, ProtocolEnum protocol, int port, string vpcId) { var createResponse = await _amazonElasticLoadBalancingV2.CreateTargetGroupAsync( new CreateTargetGroupRequest() { Name = groupName, Protocol = protocol, Port = port, HealthCheckPath = "/healthcheck", HealthCheckIntervalSeconds = 10, HealthCheckTimeoutSeconds = 5, HealthyThresholdCount = 2, UnhealthyThresholdCount = 2, VpcId = vpcId }); var targetGroup = createResponse.TargetGroups[0]; return targetGroup; } /// <summary> /// Create an Elastic Load Balancing load balancer that uses the specified subnets /// and forwards requests to the specified target group. /// </summary> /// <param name="name">The name for the new load balancer.</param> /// <param name="subnetIds">Subnets for the load balancer.</param> /// <param name="targetGroup">Target group for forwarded requests.</param> /// <returns>The new LoadBalancer object.</returns> public async Task<LoadBalancer> CreateLoadBalancerAndListener(string name, List<string> subnetIds, TargetGroup targetGroup) { var createLbResponse = await _amazonElasticLoadBalancingV2.CreateLoadBalancerAsync( new CreateLoadBalancerRequest() { Name = name, Subnets = subnetIds }); var loadBalancerArn = createLbResponse.LoadBalancers[0].LoadBalancerArn; // Wait for load balancer to be available. var loadBalancerReady = false; while (!loadBalancerReady) { try { var describeResponse = await _amazonElasticLoadBalancingV2.DescribeLoadBalancersAsync( new DescribeLoadBalancersRequest() { Names = new List<string>() { name } }); var loadBalancerState = describeResponse.LoadBalancers[0].State.Code; loadBalancerReady = loadBalancerState == LoadBalancerStateEnum.Active; } catch (LoadBalancerNotFoundException) { loadBalancerReady = false; } Thread.Sleep(10000); } // Create the listener. await _amazonElasticLoadBalancingV2.CreateListenerAsync( new CreateListenerRequest() { LoadBalancerArn = loadBalancerArn, Protocol = targetGroup.Protocol, Port = targetGroup.Port, DefaultActions = new List<Action>() { new Action() { Type = ActionTypeEnum.Forward, TargetGroupArn = targetGroup.TargetGroupArn } } }); return createLbResponse.LoadBalancers[0]; } /// <summary> /// Verify this computer can successfully send a GET request to the /// load balancer endpoint. /// </summary> /// <param name="endpoint">The endpoint to check.</param> /// <returns>True if successful.</returns> public async Task<bool> VerifyLoadBalancerEndpoint(string endpoint) { var success = false; var retries = 3; while (!success && retries > 0) { try { var endpointResponse = await _httpClient.GetAsync($"http://{endpoint}"); Console.WriteLine($"Response: {endpointResponse.StatusCode}."); if (endpointResponse.IsSuccessStatusCode) { success = true; } else { retries = 0; } } catch (HttpRequestException) { Console.WriteLine("Connection error, retrying..."); retries--; Thread.Sleep(10000); } } return success; } /// <summary> /// Delete a load balancer by its specified name. /// </summary> /// <param name="name">The name of the load balancer to delete.</param> /// <returns>Async task.</returns> public async Task DeleteLoadBalancerByName(string name) { try { var describeLoadBalancerResponse = await _amazonElasticLoadBalancingV2.DescribeLoadBalancersAsync( new DescribeLoadBalancersRequest() { Names = new List<string>() { name } }); var lbArn = describeLoadBalancerResponse.LoadBalancers[0].LoadBalancerArn; await _amazonElasticLoadBalancingV2.DeleteLoadBalancerAsync( new DeleteLoadBalancerRequest() { LoadBalancerArn = lbArn } ); } catch (LoadBalancerNotFoundException) { Console.WriteLine($"Load balancer {name} not found."); } } /// <summary> /// Delete a TargetGroup by its specified name. /// </summary> /// <param name="groupName">Name of the group to delete.</param> /// <returns>Async task.</returns> public async Task DeleteTargetGroupByName(string groupName) { var done = false; while (!done) { try { var groupResponse = await _amazonElasticLoadBalancingV2.DescribeTargetGroupsAsync( new DescribeTargetGroupsRequest() { Names = new List<string>() { groupName } }); var targetArn = groupResponse.TargetGroups[0].TargetGroupArn; await _amazonElasticLoadBalancingV2.DeleteTargetGroupAsync( new DeleteTargetGroupRequest() { TargetGroupArn = targetArn }); Console.WriteLine($"Deleted load balancing target group {groupName}."); done = true; } catch (TargetGroupNotFoundException) { Console.WriteLine( $"Target group {groupName} not found, could not delete."); done = true; } catch (ResourceInUseException) { Console.WriteLine("Target group not yet released, waiting..."); Thread.Sleep(10000); } } } }

Membuat kelas yang menggunakan DynamoDB untuk menyimulasikan layanan yang direkomendasikan.

/// <summary> /// Encapsulates a DynamoDB table to use as a service that recommends books, movies, and songs. /// </summary> public class Recommendations { private readonly IAmazonDynamoDB _amazonDynamoDb; private readonly DynamoDBContext _context; private readonly string _tableName; public string TableName => _tableName; /// <summary> /// Constructor for the Recommendations service. /// </summary> /// <param name="amazonDynamoDb">The injected DynamoDb client.</param> /// <param name="configuration">The injected configuration.</param> public Recommendations(IAmazonDynamoDB amazonDynamoDb, IConfiguration configuration) { _amazonDynamoDb = amazonDynamoDb; _context = new DynamoDBContext(_amazonDynamoDb); _tableName = configuration["databaseName"]!; } /// <summary> /// Create the DynamoDb table with a specified name. /// </summary> /// <param name="tableName">The name for the table.</param> /// <returns>True when ready.</returns> public async Task<bool> CreateDatabaseWithName(string tableName) { try { Console.Write($"Creating table {tableName}..."); var createRequest = new CreateTableRequest() { TableName = tableName, AttributeDefinitions = new List<AttributeDefinition>() { new AttributeDefinition() { AttributeName = "MediaType", AttributeType = ScalarAttributeType.S }, new AttributeDefinition() { AttributeName = "ItemId", AttributeType = ScalarAttributeType.N } }, KeySchema = new List<KeySchemaElement>() { new KeySchemaElement() { AttributeName = "MediaType", KeyType = KeyType.HASH }, new KeySchemaElement() { AttributeName = "ItemId", KeyType = KeyType.RANGE } }, ProvisionedThroughput = new ProvisionedThroughput() { ReadCapacityUnits = 5, WriteCapacityUnits = 5 } }; await _amazonDynamoDb.CreateTableAsync(createRequest); // Wait until the table is ACTIVE and then report success. Console.Write("\nWaiting for table to become active..."); var request = new DescribeTableRequest { TableName = tableName }; TableStatus status; do { Thread.Sleep(2000); var describeTableResponse = await _amazonDynamoDb.DescribeTableAsync(request); status = describeTableResponse.Table.TableStatus; Console.Write("."); } while (status != "ACTIVE"); return status == TableStatus.ACTIVE; } catch (ResourceInUseException) { Console.WriteLine($"Table {tableName} already exists."); return false; } } /// <summary> /// Populate the database table with data from a specified path. /// </summary> /// <param name="databaseTableName">The name of the table.</param> /// <param name="recommendationsPath">The path of the recommendations data.</param> /// <returns>Async task.</returns> public async Task PopulateDatabase(string databaseTableName, string recommendationsPath) { var recommendationsText = await File.ReadAllTextAsync(recommendationsPath); var records = JsonSerializer.Deserialize<RecommendationModel[]>(recommendationsText); var batchWrite = _context.CreateBatchWrite<RecommendationModel>(); foreach (var record in records!) { batchWrite.AddPutItem(record); } await batchWrite.ExecuteAsync(); } /// <summary> /// Delete the recommendation table by name. /// </summary> /// <param name="tableName">The name of the recommendation table.</param> /// <returns>Async task.</returns> public async Task DestroyDatabaseByName(string tableName) { try { await _amazonDynamoDb.DeleteTableAsync( new DeleteTableRequest() { TableName = tableName }); Console.WriteLine($"Table {tableName} was deleted."); } catch (ResourceNotFoundException) { Console.WriteLine($"Table {tableName} not found"); } } }

Membuat kelas yang mengabungkan tindakan Systems Manager.

/// <summary> /// Encapsulates Systems Manager parameter operations. This example uses these parameters /// to drive the demonstration of resilient architecture, such as failure of a dependency or /// how the service responds to a health check. /// </summary> public class SmParameterWrapper { private readonly IAmazonSimpleSystemsManagement _amazonSimpleSystemsManagement; private readonly string _tableParameter = "doc-example-resilient-architecture-table"; private readonly string _failureResponseParameter = "doc-example-resilient-architecture-failure-response"; private readonly string _healthCheckParameter = "doc-example-resilient-architecture-health-check"; private readonly string _tableName = ""; public string TableParameter => _tableParameter; public string TableName => _tableName; public string HealthCheckParameter => _healthCheckParameter; public string FailureResponseParameter => _failureResponseParameter; /// <summary> /// Constructor for the SmParameterWrapper. /// </summary> /// <param name="amazonSimpleSystemsManagement">The injected Simple Systems Management client.</param> /// <param name="configuration">The injected configuration.</param> public SmParameterWrapper(IAmazonSimpleSystemsManagement amazonSimpleSystemsManagement, IConfiguration configuration) { _amazonSimpleSystemsManagement = amazonSimpleSystemsManagement; _tableName = configuration["databaseName"]!; } /// <summary> /// Reset the Systems Manager parameters to starting values for the demo. /// </summary> /// <returns>Async task.</returns> public async Task Reset() { await this.PutParameterByName(_tableParameter, _tableName); await this.PutParameterByName(_failureResponseParameter, "none"); await this.PutParameterByName(_healthCheckParameter, "shallow"); } /// <summary> /// Set the value of a named Systems Manager parameter. /// </summary> /// <param name="name">The name of the parameter.</param> /// <param name="value">The value to set.</param> /// <returns>Async task.</returns> public async Task PutParameterByName(string name, string value) { await _amazonSimpleSystemsManagement.PutParameterAsync( new PutParameterRequest() { Name = name, Value = value, Overwrite = true }); } }
Java
SDK for Java 2.x
catatan

Ada lebih banyak tentang GitHub. Temukan contoh lengkapnya dan pelajari cara pengaturan dan menjalankannya di Repositori Contoh Kode AWS.

Menjalankan skenario interaktif di prompt perintah.

public class Main { public static final String fileName = "C:\\AWS\\resworkflow\\recommendations.json"; // Modify file location. public static final String tableName = "doc-example-recommendation-service"; public static final String startScript = "C:\\AWS\\resworkflow\\server_startup_script.sh"; // Modify file location. public static final String policyFile = "C:\\AWS\\resworkflow\\instance_policy.json"; // Modify file location. public static final String ssmJSON = "C:\\AWS\\resworkflow\\ssm_only_policy.json"; // Modify file location. public static final String failureResponse = "doc-example-resilient-architecture-failure-response"; public static final String healthCheck = "doc-example-resilient-architecture-health-check"; public static final String templateName = "doc-example-resilience-template"; public static final String roleName = "doc-example-resilience-role"; public static final String policyName = "doc-example-resilience-pol"; public static final String profileName = "doc-example-resilience-prof"; public static final String badCredsProfileName = "doc-example-resilience-prof-bc"; public static final String targetGroupName = "doc-example-resilience-tg"; public static final String autoScalingGroupName = "doc-example-resilience-group"; public static final String lbName = "doc-example-resilience-lb"; public static final String protocol = "HTTP"; public static final int port = 80; public static final String DASHES = new String(new char[80]).replace("\0", "-"); public static void main(String[] args) throws IOException, InterruptedException { Scanner in = new Scanner(System.in); Database database = new Database(); AutoScaler autoScaler = new AutoScaler(); LoadBalancer loadBalancer = new LoadBalancer(); System.out.println(DASHES); System.out.println("Welcome to the demonstration of How to Build and Manage a Resilient Service!"); System.out.println(DASHES); System.out.println(DASHES); System.out.println("A - SETUP THE RESOURCES"); System.out.println("Press Enter when you're ready to start deploying resources."); in.nextLine(); deploy(loadBalancer); System.out.println(DASHES); System.out.println(DASHES); System.out.println("B - DEMO THE RESILIENCE FUNCTIONALITY"); System.out.println("Press Enter when you're ready."); in.nextLine(); demo(loadBalancer); System.out.println(DASHES); System.out.println(DASHES); System.out.println("C - DELETE THE RESOURCES"); System.out.println(""" This concludes the demo of how to build and manage a resilient service. To keep things tidy and to avoid unwanted charges on your account, we can clean up all AWS resources that were created for this demo. """); System.out.println("\n Do you want to delete the resources (y/n)? "); String userInput = in.nextLine().trim().toLowerCase(); // Capture user input if (userInput.equals("y")) { // Delete resources here deleteResources(loadBalancer, autoScaler, database); System.out.println("Resources deleted."); } else { System.out.println(""" Okay, we'll leave the resources intact. Don't forget to delete them when you're done with them or you might incur unexpected charges. """); } System.out.println(DASHES); System.out.println(DASHES); System.out.println("The example has completed. "); System.out.println("\n Thanks for watching!"); System.out.println(DASHES); } // Deletes the AWS resources used in this example. private static void deleteResources(LoadBalancer loadBalancer, AutoScaler autoScaler, Database database) throws IOException, InterruptedException { loadBalancer.deleteLoadBalancer(lbName); System.out.println("*** Wait 30 secs for resource to be deleted"); TimeUnit.SECONDS.sleep(30); loadBalancer.deleteTargetGroup(targetGroupName); autoScaler.deleteAutoScaleGroup(autoScalingGroupName); autoScaler.deleteRolesPolicies(policyName, roleName, profileName); autoScaler.deleteTemplate(templateName); database.deleteTable(tableName); } private static void deploy(LoadBalancer loadBalancer) throws InterruptedException, IOException { Scanner in = new Scanner(System.in); System.out.println( """ For this demo, we'll use the AWS SDK for Java (v2) to create several AWS resources to set up a load-balanced web service endpoint and explore some ways to make it resilient against various kinds of failures. Some of the resources create by this demo are: \t* A DynamoDB table that the web service depends on to provide book, movie, and song recommendations. \t* An EC2 launch template that defines EC2 instances that each contain a Python web server. \t* An EC2 Auto Scaling group that manages EC2 instances across several Availability Zones. \t* An Elastic Load Balancing (ELB) load balancer that targets the Auto Scaling group to distribute requests. """); System.out.println("Press Enter when you're ready."); in.nextLine(); System.out.println(DASHES); System.out.println(DASHES); System.out.println("Creating and populating a DynamoDB table named " + tableName); Database database = new Database(); database.createTable(tableName, fileName); System.out.println(DASHES); System.out.println(DASHES); System.out.println(""" Creating an EC2 launch template that runs '{startup_script}' when an instance starts. This script starts a Python web server defined in the `server.py` script. The web server listens to HTTP requests on port 80 and responds to requests to '/' and to '/healthcheck'. For demo purposes, this server is run as the root user. In production, the best practice is to run a web server, such as Apache, with least-privileged credentials. The template also defines an IAM policy that each instance uses to assume a role that grants permissions to access the DynamoDB recommendation table and Systems Manager parameters that control the flow of the demo. """); LaunchTemplateCreator templateCreator = new LaunchTemplateCreator(); templateCreator.createTemplate(policyFile, policyName, profileName, startScript, templateName, roleName); System.out.println(DASHES); System.out.println(DASHES); System.out.println( "Creating an EC2 Auto Scaling group that maintains three EC2 instances, each in a different Availability Zone."); System.out.println("*** Wait 30 secs for the VPC to be created"); TimeUnit.SECONDS.sleep(30); AutoScaler autoScaler = new AutoScaler(); String[] zones = autoScaler.createGroup(3, templateName, autoScalingGroupName); System.out.println(""" At this point, you have EC2 instances created. Once each instance starts, it listens for HTTP requests. You can see these instances in the console or continue with the demo. Press Enter when you're ready to continue. """); in.nextLine(); System.out.println(DASHES); System.out.println(DASHES); System.out.println("Creating variables that control the flow of the demo."); ParameterHelper paramHelper = new ParameterHelper(); paramHelper.reset(); System.out.println(DASHES); System.out.println(DASHES); System.out.println(""" Creating an Elastic Load Balancing target group and load balancer. The target group defines how the load balancer connects to instances. The load balancer provides a single endpoint where clients connect and dispatches requests to instances in the group. """); String vpcId = autoScaler.getDefaultVPC(); List<Subnet> subnets = autoScaler.getSubnets(vpcId, zones); System.out.println("You have retrieved a list with " + subnets.size() + " subnets"); String targetGroupArn = loadBalancer.createTargetGroup(protocol, port, vpcId, targetGroupName); String elbDnsName = loadBalancer.createLoadBalancer(subnets, targetGroupArn, lbName, port, protocol); autoScaler.attachLoadBalancerTargetGroup(autoScalingGroupName, targetGroupArn); System.out.println("Verifying access to the load balancer endpoint..."); boolean wasSuccessul = loadBalancer.verifyLoadBalancerEndpoint(elbDnsName); if (!wasSuccessul) { System.out.println("Couldn't connect to the load balancer, verifying that the port is open..."); CloseableHttpClient httpClient = HttpClients.createDefault(); // Create an HTTP GET request to "http://checkip.amazonaws.com" HttpGet httpGet = new HttpGet("http://checkip.amazonaws.com"); try { // Execute the request and get the response HttpResponse response = httpClient.execute(httpGet); // Read the response content. String ipAddress = IOUtils.toString(response.getEntity().getContent(), StandardCharsets.UTF_8).trim(); // Print the public IP address. System.out.println("Public IP Address: " + ipAddress); GroupInfo groupInfo = autoScaler.verifyInboundPort(vpcId, port, ipAddress); if (!groupInfo.isPortOpen()) { System.out.println(""" For this example to work, the default security group for your default VPC must allow access from this computer. You can either add it automatically from this example or add it yourself using the AWS Management Console. """); System.out.println( "Do you want to add a rule to security group " + groupInfo.getGroupName() + " to allow"); System.out.println("inbound traffic on port " + port + " from your computer's IP address (y/n) "); String ans = in.nextLine(); if ("y".equalsIgnoreCase(ans)) { autoScaler.openInboundPort(groupInfo.getGroupName(), String.valueOf(port), ipAddress); System.out.println("Security group rule added."); } else { System.out.println("No security group rule added."); } } } catch (AutoScalingException e) { e.printStackTrace(); } } else if (wasSuccessul) { System.out.println("Your load balancer is ready. You can access it by browsing to:"); System.out.println("\t http://" + elbDnsName); } else { System.out.println("Couldn't get a successful response from the load balancer endpoint. Troubleshoot by"); System.out.println("manually verifying that your VPC and security group are configured correctly and that"); System.out.println("you can successfully make a GET request to the load balancer."); } System.out.println("Press Enter when you're ready to continue with the demo."); in.nextLine(); } // A method that controls the demo part of the Java program. public static void demo(LoadBalancer loadBalancer) throws IOException, InterruptedException { ParameterHelper paramHelper = new ParameterHelper(); System.out.println("Read the ssm_only_policy.json file"); String ssmOnlyPolicy = readFileAsString(ssmJSON); System.out.println("Resetting parameters to starting values for demo."); paramHelper.reset(); System.out.println( """ This part of the demonstration shows how to toggle different parts of the system to create situations where the web service fails, and shows how using a resilient architecture can keep the web service running in spite of these failures. At the start, the load balancer endpoint returns recommendations and reports that all targets are healthy. """); demoChoices(loadBalancer); System.out.println( """ The web service running on the EC2 instances gets recommendations by querying a DynamoDB table. The table name is contained in a Systems Manager parameter named self.param_helper.table. To simulate a failure of the recommendation service, let's set this parameter to name a non-existent table. """); paramHelper.put(paramHelper.tableName, "this-is-not-a-table"); System.out.println( """ \nNow, sending a GET request to the load balancer endpoint returns a failure code. But, the service reports as healthy to the load balancer because shallow health checks don't check for failure of the recommendation service. """); demoChoices(loadBalancer); System.out.println( """ Instead of failing when the recommendation service fails, the web service can return a static response. While this is not a perfect solution, it presents the customer with a somewhat better experience than failure. """); paramHelper.put(paramHelper.failureResponse, "static"); System.out.println(""" Now, sending a GET request to the load balancer endpoint returns a static response. The service still reports as healthy because health checks are still shallow. """); demoChoices(loadBalancer); System.out.println("Let's reinstate the recommendation service."); paramHelper.put(paramHelper.tableName, paramHelper.dyntable); System.out.println(""" Let's also substitute bad credentials for one of the instances in the target group so that it can't access the DynamoDB recommendation table. We will get an instance id value. """); LaunchTemplateCreator templateCreator = new LaunchTemplateCreator(); AutoScaler autoScaler = new AutoScaler(); // Create a new instance profile based on badCredsProfileName. templateCreator.createInstanceProfile(policyFile, policyName, badCredsProfileName, roleName); String badInstanceId = autoScaler.getBadInstance(autoScalingGroupName); System.out.println("The bad instance id values used for this demo is " + badInstanceId); String profileAssociationId = autoScaler.getInstanceProfile(badInstanceId); System.out.println("The association Id value is " + profileAssociationId); System.out.println("Replacing the profile for instance " + badInstanceId + " with a profile that contains bad credentials"); autoScaler.replaceInstanceProfile(badInstanceId, badCredsProfileName, profileAssociationId); System.out.println( """ Now, sending a GET request to the load balancer endpoint returns either a recommendation or a static response, depending on which instance is selected by the load balancer. """); demoChoices(loadBalancer); System.out.println(""" Let's implement a deep health check. For this demo, a deep health check tests whether the web service can access the DynamoDB table that it depends on for recommendations. Note that the deep health check is only for ELB routing and not for Auto Scaling instance health. This kind of deep health check is not recommended for Auto Scaling instance health, because it risks accidental termination of all instances in the Auto Scaling group when a dependent service fails. """); System.out.println(""" By implementing deep health checks, the load balancer can detect when one of the instances is failing and take that instance out of rotation. """); paramHelper.put(paramHelper.healthCheck, "deep"); System.out.println(""" Now, checking target health indicates that the instance with bad credentials is unhealthy. Note that it might take a minute or two for the load balancer to detect the unhealthy instance. Sending a GET request to the load balancer endpoint always returns a recommendation, because the load balancer takes unhealthy instances out of its rotation. """); demoChoices(loadBalancer); System.out.println( """ Because the instances in this demo are controlled by an auto scaler, the simplest way to fix an unhealthy instance is to terminate it and let the auto scaler start a new instance to replace it. """); autoScaler.terminateInstance(badInstanceId); System.out.println(""" Even while the instance is terminating and the new instance is starting, sending a GET request to the web service continues to get a successful recommendation response because the load balancer routes requests to the healthy instances. After the replacement instance starts and reports as healthy, it is included in the load balancing rotation. Note that terminating and replacing an instance typically takes several minutes, during which time you can see the changing health check status until the new instance is running and healthy. """); demoChoices(loadBalancer); System.out.println( "If the recommendation service fails now, deep health checks mean all instances report as unhealthy."); paramHelper.put(paramHelper.tableName, "this-is-not-a-table"); demoChoices(loadBalancer); paramHelper.reset(); } public static void demoChoices(LoadBalancer loadBalancer) throws IOException, InterruptedException { String[] actions = { "Send a GET request to the load balancer endpoint.", "Check the health of load balancer targets.", "Go to the next part of the demo." }; Scanner scanner = new Scanner(System.in); while (true) { System.out.println("-".repeat(88)); System.out.println("See the current state of the service by selecting one of the following choices:"); for (int i = 0; i < actions.length; i++) { System.out.println(i + ": " + actions[i]); } try { System.out.print("\nWhich action would you like to take? "); int choice = scanner.nextInt(); System.out.println("-".repeat(88)); switch (choice) { case 0 -> { System.out.println("Request:\n"); System.out.println("GET http://" + loadBalancer.getEndpoint(lbName)); CloseableHttpClient httpClient = HttpClients.createDefault(); // Create an HTTP GET request to the ELB. HttpGet httpGet = new HttpGet("http://" + loadBalancer.getEndpoint(lbName)); // Execute the request and get the response. HttpResponse response = httpClient.execute(httpGet); int statusCode = response.getStatusLine().getStatusCode(); System.out.println("HTTP Status Code: " + statusCode); // Display the JSON response BufferedReader reader = new BufferedReader( new InputStreamReader(response.getEntity().getContent())); StringBuilder jsonResponse = new StringBuilder(); String line; while ((line = reader.readLine()) != null) { jsonResponse.append(line); } reader.close(); // Print the formatted JSON response. System.out.println("Full Response:\n"); System.out.println(jsonResponse.toString()); // Close the HTTP client. httpClient.close(); } case 1 -> { System.out.println("\nChecking the health of load balancer targets:\n"); List<TargetHealthDescription> health = loadBalancer.checkTargetHealth(targetGroupName); for (TargetHealthDescription target : health) { System.out.printf("\tTarget %s on port %d is %s%n", target.target().id(), target.target().port(), target.targetHealth().stateAsString()); } System.out.println(""" Note that it can take a minute or two for the health check to update after changes are made. """); } case 2 -> { System.out.println("\nOkay, let's move on."); System.out.println("-".repeat(88)); return; // Exit the method when choice is 2 } default -> System.out.println("You must choose a value between 0-2. Please select again."); } } catch (java.util.InputMismatchException e) { System.out.println("Invalid input. Please select again."); scanner.nextLine(); // Clear the input buffer. } } } public static String readFileAsString(String filePath) throws IOException { byte[] bytes = Files.readAllBytes(Paths.get(filePath)); return new String(bytes); } }

Membuat kelas yang menggabungkan tindakan Penskalaan Otomatis dan Amazon EC2.

public class AutoScaler { private static Ec2Client ec2Client; private static AutoScalingClient autoScalingClient; private static IamClient iamClient; private static SsmClient ssmClient; private IamClient getIAMClient() { if (iamClient == null) { iamClient = IamClient.builder() .region(Region.US_EAST_1) .build(); } return iamClient; } private SsmClient getSSMClient() { if (ssmClient == null) { ssmClient = SsmClient.builder() .region(Region.US_EAST_1) .build(); } return ssmClient; } private Ec2Client getEc2Client() { if (ec2Client == null) { ec2Client = Ec2Client.builder() .region(Region.US_EAST_1) .build(); } return ec2Client; } private AutoScalingClient getAutoScalingClient() { if (autoScalingClient == null) { autoScalingClient = AutoScalingClient.builder() .region(Region.US_EAST_1) .build(); } return autoScalingClient; } /** * Terminates and instances in an EC2 Auto Scaling group. After an instance is * terminated, it can no longer be accessed. */ public void terminateInstance(String instanceId) { TerminateInstanceInAutoScalingGroupRequest terminateInstanceIRequest = TerminateInstanceInAutoScalingGroupRequest .builder() .instanceId(instanceId) .shouldDecrementDesiredCapacity(false) .build(); getAutoScalingClient().terminateInstanceInAutoScalingGroup(terminateInstanceIRequest); System.out.format("Terminated instance %s.", instanceId); } /** * Replaces the profile associated with a running instance. After the profile is * replaced, the instance is rebooted to ensure that it uses the new profile. * When * the instance is ready, Systems Manager is used to restart the Python web * server. */ public void replaceInstanceProfile(String instanceId, String newInstanceProfileName, String profileAssociationId) throws InterruptedException { // Create an IAM instance profile specification. software.amazon.awssdk.services.ec2.model.IamInstanceProfileSpecification iamInstanceProfile = software.amazon.awssdk.services.ec2.model.IamInstanceProfileSpecification .builder() .name(newInstanceProfileName) // Make sure 'newInstanceProfileName' is a valid IAM Instance Profile // name. .build(); // Replace the IAM instance profile association for the EC2 instance. ReplaceIamInstanceProfileAssociationRequest replaceRequest = ReplaceIamInstanceProfileAssociationRequest .builder() .iamInstanceProfile(iamInstanceProfile) .associationId(profileAssociationId) // Make sure 'profileAssociationId' is a valid association ID. .build(); try { getEc2Client().replaceIamInstanceProfileAssociation(replaceRequest); // Handle the response as needed. } catch (Ec2Exception e) { // Handle exceptions, log, or report the error. System.err.println("Error: " + e.getMessage()); } System.out.format("Replaced instance profile for association %s with profile %s.", profileAssociationId, newInstanceProfileName); TimeUnit.SECONDS.sleep(15); boolean instReady = false; int tries = 0; // Reboot after 60 seconds while (!instReady) { if (tries % 6 == 0) { getEc2Client().rebootInstances(RebootInstancesRequest.builder() .instanceIds(instanceId) .build()); System.out.println("Rebooting instance " + instanceId + " and waiting for it to be ready."); } tries++; try { TimeUnit.SECONDS.sleep(10); } catch (InterruptedException e) { e.printStackTrace(); } DescribeInstanceInformationResponse informationResponse = getSSMClient().describeInstanceInformation(); List<InstanceInformation> instanceInformationList = informationResponse.instanceInformationList(); for (InstanceInformation info : instanceInformationList) { if (info.instanceId().equals(instanceId)) { instReady = true; break; } } } SendCommandRequest sendCommandRequest = SendCommandRequest.builder() .instanceIds(instanceId) .documentName("AWS-RunShellScript") .parameters(Collections.singletonMap("commands", Collections.singletonList("cd / && sudo python3 server.py 80"))) .build(); getSSMClient().sendCommand(sendCommandRequest); System.out.println("Restarted the Python web server on instance " + instanceId + "."); } public void openInboundPort(String secGroupId, String port, String ipAddress) { AuthorizeSecurityGroupIngressRequest ingressRequest = AuthorizeSecurityGroupIngressRequest.builder() .groupName(secGroupId) .cidrIp(ipAddress) .fromPort(Integer.parseInt(port)) .build(); getEc2Client().authorizeSecurityGroupIngress(ingressRequest); System.out.format("Authorized ingress to %s on port %s from %s.", secGroupId, port, ipAddress); } /** * Detaches a role from an instance profile, detaches policies from the role, * and deletes all the resources. */ public void deleteInstanceProfile(String roleName, String profileName) { try { software.amazon.awssdk.services.iam.model.GetInstanceProfileRequest getInstanceProfileRequest = software.amazon.awssdk.services.iam.model.GetInstanceProfileRequest .builder() .instanceProfileName(profileName) .build(); GetInstanceProfileResponse response = getIAMClient().getInstanceProfile(getInstanceProfileRequest); String name = response.instanceProfile().instanceProfileName(); System.out.println(name); RemoveRoleFromInstanceProfileRequest profileRequest = RemoveRoleFromInstanceProfileRequest.builder() .instanceProfileName(profileName) .roleName(roleName) .build(); getIAMClient().removeRoleFromInstanceProfile(profileRequest); DeleteInstanceProfileRequest deleteInstanceProfileRequest = DeleteInstanceProfileRequest.builder() .instanceProfileName(profileName) .build(); getIAMClient().deleteInstanceProfile(deleteInstanceProfileRequest); System.out.println("Deleted instance profile " + profileName); DeleteRoleRequest deleteRoleRequest = DeleteRoleRequest.builder() .roleName(roleName) .build(); // List attached role policies. ListAttachedRolePoliciesResponse rolesResponse = getIAMClient() .listAttachedRolePolicies(role -> role.roleName(roleName)); List<AttachedPolicy> attachedPolicies = rolesResponse.attachedPolicies(); for (AttachedPolicy attachedPolicy : attachedPolicies) { DetachRolePolicyRequest request = DetachRolePolicyRequest.builder() .roleName(roleName) .policyArn(attachedPolicy.policyArn()) .build(); getIAMClient().detachRolePolicy(request); System.out.println("Detached and deleted policy " + attachedPolicy.policyName()); } getIAMClient().deleteRole(deleteRoleRequest); System.out.println("Instance profile and role deleted."); } catch (IamException e) { System.err.println(e.getMessage()); System.exit(1); } } public void deleteTemplate(String templateName) { getEc2Client().deleteLaunchTemplate(name -> name.launchTemplateName(templateName)); System.out.format(templateName + " was deleted."); } public void deleteAutoScaleGroup(String groupName) { DeleteAutoScalingGroupRequest deleteAutoScalingGroupRequest = DeleteAutoScalingGroupRequest.builder() .autoScalingGroupName(groupName) .forceDelete(true) .build(); getAutoScalingClient().deleteAutoScalingGroup(deleteAutoScalingGroupRequest); System.out.println(groupName + " was deleted."); } /* * Verify the default security group of the specified VPC allows ingress from * this * computer. This can be done by allowing ingress from this computer's IP * address. In some situations, such as connecting from a corporate network, you * must instead specify a prefix list ID. You can also temporarily open the port * to * any IP address while running this example. If you do, be sure to remove * public * access when you're done. * */ public GroupInfo verifyInboundPort(String VPC, int port, String ipAddress) { boolean portIsOpen = false; GroupInfo groupInfo = new GroupInfo(); try { Filter filter = Filter.builder() .name("group-name") .values("default") .build(); Filter filter1 = Filter.builder() .name("vpc-id") .values(VPC) .build(); DescribeSecurityGroupsRequest securityGroupsRequest = DescribeSecurityGroupsRequest.builder() .filters(filter, filter1) .build(); DescribeSecurityGroupsResponse securityGroupsResponse = getEc2Client() .describeSecurityGroups(securityGroupsRequest); String securityGroup = securityGroupsResponse.securityGroups().get(0).groupName(); groupInfo.setGroupName(securityGroup); for (SecurityGroup secGroup : securityGroupsResponse.securityGroups()) { System.out.println("Found security group: " + secGroup.groupId()); for (IpPermission ipPermission : secGroup.ipPermissions()) { if (ipPermission.fromPort() == port) { System.out.println("Found inbound rule: " + ipPermission); for (IpRange ipRange : ipPermission.ipRanges()) { String cidrIp = ipRange.cidrIp(); if (cidrIp.startsWith(ipAddress) || cidrIp.equals("0.0.0.0/0")) { System.out.println(cidrIp + " is applicable"); portIsOpen = true; } } if (!ipPermission.prefixListIds().isEmpty()) { System.out.println("Prefix lList is applicable"); portIsOpen = true; } if (!portIsOpen) { System.out .println("The inbound rule does not appear to be open to either this computer's IP," + " all IP addresses (0.0.0.0/0), or to a prefix list ID."); } else { break; } } } } } catch (AutoScalingException e) { System.err.println(e.awsErrorDetails().errorMessage()); } groupInfo.setPortOpen(portIsOpen); return groupInfo; } /* * Attaches an Elastic Load Balancing (ELB) target group to this EC2 Auto * Scaling group. * The target group specifies how the load balancer forward requests to the * instances * in the group. */ public void attachLoadBalancerTargetGroup(String asGroupName, String targetGroupARN) { try { AttachLoadBalancerTargetGroupsRequest targetGroupsRequest = AttachLoadBalancerTargetGroupsRequest.builder() .autoScalingGroupName(asGroupName) .targetGroupARNs(targetGroupARN) .build(); getAutoScalingClient().attachLoadBalancerTargetGroups(targetGroupsRequest); System.out.println("Attached load balancer to " + asGroupName); } catch (AutoScalingException e) { System.err.println(e.awsErrorDetails().errorMessage()); System.exit(1); } } // Creates an EC2 Auto Scaling group with the specified size. public String[] createGroup(int groupSize, String templateName, String autoScalingGroupName) { // Get availability zones. software.amazon.awssdk.services.ec2.model.DescribeAvailabilityZonesRequest zonesRequest = software.amazon.awssdk.services.ec2.model.DescribeAvailabilityZonesRequest .builder() .build(); DescribeAvailabilityZonesResponse zonesResponse = getEc2Client().describeAvailabilityZones(zonesRequest); List<String> availabilityZoneNames = zonesResponse.availabilityZones().stream() .map(software.amazon.awssdk.services.ec2.model.AvailabilityZone::zoneName) .collect(Collectors.toList()); String availabilityZones = String.join(",", availabilityZoneNames); LaunchTemplateSpecification specification = LaunchTemplateSpecification.builder() .launchTemplateName(templateName) .version("$Default") .build(); String[] zones = availabilityZones.split(","); CreateAutoScalingGroupRequest groupRequest = CreateAutoScalingGroupRequest.builder() .launchTemplate(specification) .availabilityZones(zones) .maxSize(groupSize) .minSize(groupSize) .autoScalingGroupName(autoScalingGroupName) .build(); try { getAutoScalingClient().createAutoScalingGroup(groupRequest); } catch (AutoScalingException e) { System.err.println(e.awsErrorDetails().errorMessage()); System.exit(1); } System.out.println("Created an EC2 Auto Scaling group named " + autoScalingGroupName); return zones; } public String getDefaultVPC() { // Define the filter. Filter defaultFilter = Filter.builder() .name("is-default") .values("true") .build(); software.amazon.awssdk.services.ec2.model.DescribeVpcsRequest request = software.amazon.awssdk.services.ec2.model.DescribeVpcsRequest .builder() .filters(defaultFilter) .build(); DescribeVpcsResponse response = getEc2Client().describeVpcs(request); return response.vpcs().get(0).vpcId(); } // Gets the default subnets in a VPC for a specified list of Availability Zones. public List<Subnet> getSubnets(String vpcId, String[] availabilityZones) { List<Subnet> subnets = null; Filter vpcFilter = Filter.builder() .name("vpc-id") .values(vpcId) .build(); Filter azFilter = Filter.builder() .name("availability-zone") .values(availabilityZones) .build(); Filter defaultForAZ = Filter.builder() .name("default-for-az") .values("true") .build(); DescribeSubnetsRequest request = DescribeSubnetsRequest.builder() .filters(vpcFilter, azFilter, defaultForAZ) .build(); DescribeSubnetsResponse response = getEc2Client().describeSubnets(request); subnets = response.subnets(); return subnets; } // Gets data about the instances in the EC2 Auto Scaling group. public String getBadInstance(String groupName) { DescribeAutoScalingGroupsRequest request = DescribeAutoScalingGroupsRequest.builder() .autoScalingGroupNames(groupName) .build(); DescribeAutoScalingGroupsResponse response = getAutoScalingClient().describeAutoScalingGroups(request); AutoScalingGroup autoScalingGroup = response.autoScalingGroups().get(0); List<String> instanceIds = autoScalingGroup.instances().stream() .map(instance -> instance.instanceId()) .collect(Collectors.toList()); String[] instanceIdArray = instanceIds.toArray(new String[0]); for (String instanceId : instanceIdArray) { System.out.println("Instance ID: " + instanceId); return instanceId; } return ""; } // Gets data about the profile associated with an instance. public String getInstanceProfile(String instanceId) { Filter filter = Filter.builder() .name("instance-id") .values(instanceId) .build(); DescribeIamInstanceProfileAssociationsRequest associationsRequest = DescribeIamInstanceProfileAssociationsRequest .builder() .filters(filter) .build(); DescribeIamInstanceProfileAssociationsResponse response = getEc2Client() .describeIamInstanceProfileAssociations(associationsRequest); return response.iamInstanceProfileAssociations().get(0).associationId(); } public void deleteRolesPolicies(String policyName, String roleName, String InstanceProfile) { ListPoliciesRequest listPoliciesRequest = ListPoliciesRequest.builder().build(); ListPoliciesResponse listPoliciesResponse = getIAMClient().listPolicies(listPoliciesRequest); for (Policy policy : listPoliciesResponse.policies()) { if (policy.policyName().equals(policyName)) { // List the entities (users, groups, roles) that are attached to the policy. software.amazon.awssdk.services.iam.model.ListEntitiesForPolicyRequest listEntitiesRequest = software.amazon.awssdk.services.iam.model.ListEntitiesForPolicyRequest .builder() .policyArn(policy.arn()) .build(); ListEntitiesForPolicyResponse listEntitiesResponse = iamClient .listEntitiesForPolicy(listEntitiesRequest); if (!listEntitiesResponse.policyGroups().isEmpty() || !listEntitiesResponse.policyUsers().isEmpty() || !listEntitiesResponse.policyRoles().isEmpty()) { // Detach the policy from any entities it is attached to. DetachRolePolicyRequest detachPolicyRequest = DetachRolePolicyRequest.builder() .policyArn(policy.arn()) .roleName(roleName) // Specify the name of the IAM role .build(); getIAMClient().detachRolePolicy(detachPolicyRequest); System.out.println("Policy detached from entities."); } // Now, you can delete the policy. DeletePolicyRequest deletePolicyRequest = DeletePolicyRequest.builder() .policyArn(policy.arn()) .build(); getIAMClient().deletePolicy(deletePolicyRequest); System.out.println("Policy deleted successfully."); break; } } // List the roles associated with the instance profile ListInstanceProfilesForRoleRequest listRolesRequest = ListInstanceProfilesForRoleRequest.builder() .roleName(roleName) .build(); // Detach the roles from the instance profile ListInstanceProfilesForRoleResponse listRolesResponse = iamClient.listInstanceProfilesForRole(listRolesRequest); for (software.amazon.awssdk.services.iam.model.InstanceProfile profile : listRolesResponse.instanceProfiles()) { RemoveRoleFromInstanceProfileRequest removeRoleRequest = RemoveRoleFromInstanceProfileRequest.builder() .instanceProfileName(InstanceProfile) .roleName(roleName) // Remove the extra dot here .build(); getIAMClient().removeRoleFromInstanceProfile(removeRoleRequest); System.out.println("Role " + roleName + " removed from instance profile " + InstanceProfile); } // Delete the instance profile after removing all roles DeleteInstanceProfileRequest deleteInstanceProfileRequest = DeleteInstanceProfileRequest.builder() .instanceProfileName(InstanceProfile) .build(); getIAMClient().deleteInstanceProfile(r -> r.instanceProfileName(InstanceProfile)); System.out.println(InstanceProfile + " Deleted"); System.out.println("All roles and policies are deleted."); } }

Membuat kelas yang menggabungkan tindakan Penyeimbangan Beban Elastis.

public class LoadBalancer { public ElasticLoadBalancingV2Client elasticLoadBalancingV2Client; public ElasticLoadBalancingV2Client getLoadBalancerClient() { if (elasticLoadBalancingV2Client == null) { elasticLoadBalancingV2Client = ElasticLoadBalancingV2Client.builder() .region(Region.US_EAST_1) .build(); } return elasticLoadBalancingV2Client; } // Checks the health of the instances in the target group. public List<TargetHealthDescription> checkTargetHealth(String targetGroupName) { DescribeTargetGroupsRequest targetGroupsRequest = DescribeTargetGroupsRequest.builder() .names(targetGroupName) .build(); DescribeTargetGroupsResponse tgResponse = getLoadBalancerClient().describeTargetGroups(targetGroupsRequest); DescribeTargetHealthRequest healthRequest = DescribeTargetHealthRequest.builder() .targetGroupArn(tgResponse.targetGroups().get(0).targetGroupArn()) .build(); DescribeTargetHealthResponse healthResponse = getLoadBalancerClient().describeTargetHealth(healthRequest); return healthResponse.targetHealthDescriptions(); } // Gets the HTTP endpoint of the load balancer. public String getEndpoint(String lbName) { DescribeLoadBalancersResponse res = getLoadBalancerClient() .describeLoadBalancers(describe -> describe.names(lbName)); return res.loadBalancers().get(0).dnsName(); } // Deletes a load balancer. public void deleteLoadBalancer(String lbName) { try { // Use a waiter to delete the Load Balancer. DescribeLoadBalancersResponse res = getLoadBalancerClient() .describeLoadBalancers(describe -> describe.names(lbName)); ElasticLoadBalancingV2Waiter loadBalancerWaiter = getLoadBalancerClient().waiter(); DescribeLoadBalancersRequest request = DescribeLoadBalancersRequest.builder() .loadBalancerArns(res.loadBalancers().get(0).loadBalancerArn()) .build(); getLoadBalancerClient().deleteLoadBalancer( builder -> builder.loadBalancerArn(res.loadBalancers().get(0).loadBalancerArn())); WaiterResponse<DescribeLoadBalancersResponse> waiterResponse = loadBalancerWaiter .waitUntilLoadBalancersDeleted(request); waiterResponse.matched().response().ifPresent(System.out::println); } catch (ElasticLoadBalancingV2Exception e) { System.err.println(e.awsErrorDetails().errorMessage()); } System.out.println(lbName + " was deleted."); } // Deletes the target group. public void deleteTargetGroup(String targetGroupName) { try { DescribeTargetGroupsResponse res = getLoadBalancerClient() .describeTargetGroups(describe -> describe.names(targetGroupName)); getLoadBalancerClient() .deleteTargetGroup(builder -> builder.targetGroupArn(res.targetGroups().get(0).targetGroupArn())); } catch (ElasticLoadBalancingV2Exception e) { System.err.println(e.awsErrorDetails().errorMessage()); } System.out.println(targetGroupName + " was deleted."); } // Verify this computer can successfully send a GET request to the load balancer // endpoint. public boolean verifyLoadBalancerEndpoint(String elbDnsName) throws IOException, InterruptedException { boolean success = false; int retries = 3; CloseableHttpClient httpClient = HttpClients.createDefault(); // Create an HTTP GET request to the ELB. HttpGet httpGet = new HttpGet("http://" + elbDnsName); try { while ((!success) && (retries > 0)) { // Execute the request and get the response. HttpResponse response = httpClient.execute(httpGet); int statusCode = response.getStatusLine().getStatusCode(); System.out.println("HTTP Status Code: " + statusCode); if (statusCode == 200) { success = true; } else { retries--; System.out.println("Got connection error from load balancer endpoint, retrying..."); TimeUnit.SECONDS.sleep(15); } } } catch (org.apache.http.conn.HttpHostConnectException e) { System.out.println(e.getMessage()); } System.out.println("Status.." + success); return success; } /* * Creates an Elastic Load Balancing target group. The target group specifies * how * the load balancer forward requests to instances in the group and how instance * health is checked. */ public String createTargetGroup(String protocol, int port, String vpcId, String targetGroupName) { CreateTargetGroupRequest targetGroupRequest = CreateTargetGroupRequest.builder() .healthCheckPath("/healthcheck") .healthCheckTimeoutSeconds(5) .port(port) .vpcId(vpcId) .name(targetGroupName) .protocol(protocol) .build(); CreateTargetGroupResponse targetGroupResponse = getLoadBalancerClient().createTargetGroup(targetGroupRequest); String targetGroupArn = targetGroupResponse.targetGroups().get(0).targetGroupArn(); String targetGroup = targetGroupResponse.targetGroups().get(0).targetGroupName(); System.out.println("The " + targetGroup + " was created with ARN" + targetGroupArn); return targetGroupArn; } /* * Creates an Elastic Load Balancing load balancer that uses the specified * subnets * and forwards requests to the specified target group. */ public String createLoadBalancer(List<Subnet> subnetIds, String targetGroupARN, String lbName, int port, String protocol) { try { List<String> subnetIdStrings = subnetIds.stream() .map(Subnet::subnetId) .collect(Collectors.toList()); CreateLoadBalancerRequest balancerRequest = CreateLoadBalancerRequest.builder() .subnets(subnetIdStrings) .name(lbName) .scheme("internet-facing") .build(); // Create and wait for the load balancer to become available. CreateLoadBalancerResponse lsResponse = getLoadBalancerClient().createLoadBalancer(balancerRequest); String lbARN = lsResponse.loadBalancers().get(0).loadBalancerArn(); ElasticLoadBalancingV2Waiter loadBalancerWaiter = getLoadBalancerClient().waiter(); DescribeLoadBalancersRequest request = DescribeLoadBalancersRequest.builder() .loadBalancerArns(lbARN) .build(); System.out.println("Waiting for Load Balancer " + lbName + " to become available."); WaiterResponse<DescribeLoadBalancersResponse> waiterResponse = loadBalancerWaiter .waitUntilLoadBalancerAvailable(request); waiterResponse.matched().response().ifPresent(System.out::println); System.out.println("Load Balancer " + lbName + " is available."); // Get the DNS name (endpoint) of the load balancer. String lbDNSName = lsResponse.loadBalancers().get(0).dnsName(); System.out.println("*** Load Balancer DNS Name: " + lbDNSName); // Create a listener for the load balance. Action action = Action.builder() .targetGroupArn(targetGroupARN) .type("forward") .build(); CreateListenerRequest listenerRequest = CreateListenerRequest.builder() .loadBalancerArn(lsResponse.loadBalancers().get(0).loadBalancerArn()) .defaultActions(action) .port(port) .protocol(protocol) .defaultActions(action) .build(); getLoadBalancerClient().createListener(listenerRequest); System.out.println("Created listener to forward traffic from load balancer " + lbName + " to target group " + targetGroupARN); // Return the load balancer DNS name. return lbDNSName; } catch (ElasticLoadBalancingV2Exception e) { e.printStackTrace(); } return ""; } }

Membuat kelas yang menggunakan DynamoDB untuk menyimulasikan layanan yang direkomendasikan.

public class Database { private static DynamoDbClient dynamoDbClient; public static DynamoDbClient getDynamoDbClient() { if (dynamoDbClient == null) { dynamoDbClient = DynamoDbClient.builder() .region(Region.US_EAST_1) .build(); } return dynamoDbClient; } // Checks to see if the Amazon DynamoDB table exists. private boolean doesTableExist(String tableName) { try { // Describe the table and catch any exceptions. DescribeTableRequest describeTableRequest = DescribeTableRequest.builder() .tableName(tableName) .build(); getDynamoDbClient().describeTable(describeTableRequest); System.out.println("Table '" + tableName + "' exists."); return true; } catch (ResourceNotFoundException e) { System.out.println("Table '" + tableName + "' does not exist."); } catch (DynamoDbException e) { System.err.println("Error checking table existence: " + e.getMessage()); } return false; } /* * Creates a DynamoDB table to use a recommendation service. The table has a * hash key named 'MediaType' that defines the type of media recommended, such * as * Book or Movie, and a range key named 'ItemId' that, combined with the * MediaType, * forms a unique identifier for the recommended item. */ public void createTable(String tableName, String fileName) throws IOException { // First check to see if the table exists. boolean doesExist = doesTableExist(tableName); if (!doesExist) { DynamoDbWaiter dbWaiter = getDynamoDbClient().waiter(); CreateTableRequest createTableRequest = CreateTableRequest.builder() .tableName(tableName) .attributeDefinitions( AttributeDefinition.builder() .attributeName("MediaType") .attributeType(ScalarAttributeType.S) .build(), AttributeDefinition.builder() .attributeName("ItemId") .attributeType(ScalarAttributeType.N) .build()) .keySchema( KeySchemaElement.builder() .attributeName("MediaType") .keyType(KeyType.HASH) .build(), KeySchemaElement.builder() .attributeName("ItemId") .keyType(KeyType.RANGE) .build()) .provisionedThroughput( ProvisionedThroughput.builder() .readCapacityUnits(5L) .writeCapacityUnits(5L) .build()) .build(); getDynamoDbClient().createTable(createTableRequest); System.out.println("Creating table " + tableName + "..."); // Wait until the Amazon DynamoDB table is created. DescribeTableRequest tableRequest = DescribeTableRequest.builder() .tableName(tableName) .build(); WaiterResponse<DescribeTableResponse> waiterResponse = dbWaiter.waitUntilTableExists(tableRequest); waiterResponse.matched().response().ifPresent(System.out::println); System.out.println("Table " + tableName + " created."); // Add records to the table. populateTable(fileName, tableName); } } public void deleteTable(String tableName) { getDynamoDbClient().deleteTable(table -> table.tableName(tableName)); System.out.println("Table " + tableName + " deleted."); } // Populates the table with data located in a JSON file using the DynamoDB // enhanced client. public void populateTable(String fileName, String tableName) throws IOException { DynamoDbEnhancedClient enhancedClient = DynamoDbEnhancedClient.builder() .dynamoDbClient(getDynamoDbClient()) .build(); ObjectMapper objectMapper = new ObjectMapper(); File jsonFile = new File(fileName); JsonNode rootNode = objectMapper.readTree(jsonFile); DynamoDbTable<Recommendation> mappedTable = enhancedClient.table(tableName, TableSchema.fromBean(Recommendation.class)); for (JsonNode currentNode : rootNode) { String mediaType = currentNode.path("MediaType").path("S").asText(); int itemId = currentNode.path("ItemId").path("N").asInt(); String title = currentNode.path("Title").path("S").asText(); String creator = currentNode.path("Creator").path("S").asText(); // Create a Recommendation object and set its properties. Recommendation rec = new Recommendation(); rec.setMediaType(mediaType); rec.setItemId(itemId); rec.setTitle(title); rec.setCreator(creator); // Put the item into the DynamoDB table. mappedTable.putItem(rec); // Add the Recommendation to the list. } System.out.println("Added all records to the " + tableName); } }

Membuat kelas yang mengabungkan tindakan Systems Manager.

public class ParameterHelper { String tableName = "doc-example-resilient-architecture-table"; String dyntable = "doc-example-recommendation-service"; String failureResponse = "doc-example-resilient-architecture-failure-response"; String healthCheck = "doc-example-resilient-architecture-health-check"; public void reset() { put(dyntable, tableName); put(failureResponse, "none"); put(healthCheck, "shallow"); } public void put(String name, String value) { SsmClient ssmClient = SsmClient.builder() .region(Region.US_EAST_1) .build(); PutParameterRequest parameterRequest = PutParameterRequest.builder() .name(name) .value(value) .overwrite(true) .type("String") .build(); ssmClient.putParameter(parameterRequest); System.out.printf("Setting demo parameter %s to '%s'.", name, value); } }
JavaScript
SDK untuk JavaScript (v3)
catatan

Ada lebih banyak tentang GitHub. Temukan contoh lengkapnya dan pelajari cara pengaturan dan menjalankannya di Repositori Contoh Kode AWS.

Menjalankan skenario interaktif di prompt perintah.

#!/usr/bin/env node // Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved. // SPDX-License-Identifier: Apache-2.0 // Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved. // SPDX-License-Identifier: Apache-2.0 import { Scenario, parseScenarioArgs, } from "@aws-doc-sdk-examples/lib/scenario/index.js"; /** * The workflow steps are split into three stages: * - deploy * - demo * - destroy * * Each of these stages has a corresponding file prefixed with steps-*. */ import { deploySteps } from "./steps-deploy.js"; import { demoSteps } from "./steps-demo.js"; import { destroySteps } from "./steps-destroy.js"; /** * The context is passed to every scenario. Scenario steps * will modify the context. */ const context = {}; /** * Three Scenarios are created for the workflow. A Scenario is an orchestration class * that simplifies running a series of steps. */ export const scenarios = { // Deploys all resources necessary for the workflow. deploy: new Scenario("Resilient Workflow - Deploy", deploySteps, context), // Demonstrates how a fragile web service can be made more resilient. demo: new Scenario("Resilient Workflow - Demo", demoSteps, context), // Destroys the resources created for the workflow. destroy: new Scenario("Resilient Workflow - Destroy", destroySteps, context), }; // Call function if run directly import { fileURLToPath } from "url"; if (process.argv[1] === fileURLToPath(import.meta.url)) { parseScenarioArgs(scenarios); }

Menyusun langkah-langkah untuk men-deploy semua sumber daya.

// Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved. // SPDX-License-Identifier: Apache-2.0 import { join } from "node:path"; import { readFileSync, writeFileSync } from "node:fs"; import axios from "axios"; import { BatchWriteItemCommand, CreateTableCommand, DynamoDBClient, waitUntilTableExists, } from "@aws-sdk/client-dynamodb"; import { EC2Client, CreateKeyPairCommand, CreateLaunchTemplateCommand, DescribeAvailabilityZonesCommand, DescribeVpcsCommand, DescribeSubnetsCommand, DescribeSecurityGroupsCommand, AuthorizeSecurityGroupIngressCommand, } from "@aws-sdk/client-ec2"; import { IAMClient, CreatePolicyCommand, CreateRoleCommand, CreateInstanceProfileCommand, AddRoleToInstanceProfileCommand, AttachRolePolicyCommand, waitUntilInstanceProfileExists, } from "@aws-sdk/client-iam"; import { SSMClient, GetParameterCommand } from "@aws-sdk/client-ssm"; import { CreateAutoScalingGroupCommand, AutoScalingClient, AttachLoadBalancerTargetGroupsCommand, } from "@aws-sdk/client-auto-scaling"; import { CreateListenerCommand, CreateLoadBalancerCommand, CreateTargetGroupCommand, ElasticLoadBalancingV2Client, waitUntilLoadBalancerAvailable, } from "@aws-sdk/client-elastic-load-balancing-v2"; import { ScenarioOutput, ScenarioInput, ScenarioAction, } from "@aws-doc-sdk-examples/lib/scenario/index.js"; import { retry } from "@aws-doc-sdk-examples/lib/utils/util-timers.js"; import { MESSAGES, NAMES, RESOURCES_PATH, ROOT } from "./constants.js"; import { initParamsSteps } from "./steps-reset-params.js"; /** * @type {import('@aws-doc-sdk-examples/lib/scenario.js').Step[]} */ export const deploySteps = [ new ScenarioOutput("introduction", MESSAGES.introduction, { header: true }), new ScenarioInput("confirmDeployment", MESSAGES.confirmDeployment, { type: "confirm", }), new ScenarioAction( "handleConfirmDeployment", (c) => c.confirmDeployment === false && process.exit(), ), new ScenarioOutput( "creatingTable", MESSAGES.creatingTable.replace("${TABLE_NAME}", NAMES.tableName), ), new ScenarioAction("createTable", async () => { const client = new DynamoDBClient({}); await client.send( new CreateTableCommand({ TableName: NAMES.tableName, ProvisionedThroughput: { ReadCapacityUnits: 5, WriteCapacityUnits: 5, }, AttributeDefinitions: [ { AttributeName: "MediaType", AttributeType: "S", }, { AttributeName: "ItemId", AttributeType: "N", }, ], KeySchema: [ { AttributeName: "MediaType", KeyType: "HASH", }, { AttributeName: "ItemId", KeyType: "RANGE", }, ], }), ); await waitUntilTableExists({ client }, { TableName: NAMES.tableName }); }), new ScenarioOutput( "createdTable", MESSAGES.createdTable.replace("${TABLE_NAME}", NAMES.tableName), ), new ScenarioOutput( "populatingTable", MESSAGES.populatingTable.replace("${TABLE_NAME}", NAMES.tableName), ), new ScenarioAction("populateTable", () => { const client = new DynamoDBClient({}); /** * @type {{ default: import("@aws-sdk/client-dynamodb").PutRequest['Item'][] }} */ const recommendations = JSON.parse( readFileSync(join(RESOURCES_PATH, "recommendations.json")), ); return client.send( new BatchWriteItemCommand({ RequestItems: { [NAMES.tableName]: recommendations.map((item) => ({ PutRequest: { Item: item }, })), }, }), ); }), new ScenarioOutput( "populatedTable", MESSAGES.populatedTable.replace("${TABLE_NAME}", NAMES.tableName), ), new ScenarioOutput( "creatingKeyPair", MESSAGES.creatingKeyPair.replace("${KEY_PAIR_NAME}", NAMES.keyPairName), ), new ScenarioAction("createKeyPair", async () => { const client = new EC2Client({}); const { KeyMaterial } = await client.send( new CreateKeyPairCommand({ KeyName: NAMES.keyPairName, }), ); writeFileSync(`${NAMES.keyPairName}.pem`, KeyMaterial, { mode: 0o600 }); }), new ScenarioOutput( "createdKeyPair", MESSAGES.createdKeyPair.replace("${KEY_PAIR_NAME}", NAMES.keyPairName), ), new ScenarioOutput( "creatingInstancePolicy", MESSAGES.creatingInstancePolicy.replace( "${INSTANCE_POLICY_NAME}", NAMES.instancePolicyName, ), ), new ScenarioAction("createInstancePolicy", async (state) => { const client = new IAMClient({}); const { Policy: { Arn }, } = await client.send( new CreatePolicyCommand({ PolicyName: NAMES.instancePolicyName, PolicyDocument: readFileSync( join(RESOURCES_PATH, "instance_policy.json"), ), }), ); state.instancePolicyArn = Arn; }), new ScenarioOutput("createdInstancePolicy", (state) => MESSAGES.createdInstancePolicy .replace("${INSTANCE_POLICY_NAME}", NAMES.instancePolicyName) .replace("${INSTANCE_POLICY_ARN}", state.instancePolicyArn), ), new ScenarioOutput( "creatingInstanceRole", MESSAGES.creatingInstanceRole.replace( "${INSTANCE_ROLE_NAME}", NAMES.instanceRoleName, ), ), new ScenarioAction("createInstanceRole", () => { const client = new IAMClient({}); return client.send( new CreateRoleCommand({ RoleName: NAMES.instanceRoleName, AssumeRolePolicyDocument: readFileSync( join(ROOT, "assume-role-policy.json"), ), }), ); }), new ScenarioOutput( "createdInstanceRole", MESSAGES.createdInstanceRole.replace( "${INSTANCE_ROLE_NAME}", NAMES.instanceRoleName, ), ), new ScenarioOutput( "attachingPolicyToRole", MESSAGES.attachingPolicyToRole .replace("${INSTANCE_ROLE_NAME}", NAMES.instanceRoleName) .replace("${INSTANCE_POLICY_NAME}", NAMES.instancePolicyName), ), new ScenarioAction("attachPolicyToRole", async (state) => { const client = new IAMClient({}); await client.send( new AttachRolePolicyCommand({ RoleName: NAMES.instanceRoleName, PolicyArn: state.instancePolicyArn, }), ); }), new ScenarioOutput( "attachedPolicyToRole", MESSAGES.attachedPolicyToRole .replace("${INSTANCE_POLICY_NAME}", NAMES.instancePolicyName) .replace("${INSTANCE_ROLE_NAME}", NAMES.instanceRoleName), ), new ScenarioOutput( "creatingInstanceProfile", MESSAGES.creatingInstanceProfile.replace( "${INSTANCE_PROFILE_NAME}", NAMES.instanceProfileName, ), ), new ScenarioAction("createInstanceProfile", async (state) => { const client = new IAMClient({}); const { InstanceProfile: { Arn }, } = await client.send( new CreateInstanceProfileCommand({ InstanceProfileName: NAMES.instanceProfileName, }), ); state.instanceProfileArn = Arn; await waitUntilInstanceProfileExists( { client }, { InstanceProfileName: NAMES.instanceProfileName }, ); }), new ScenarioOutput("createdInstanceProfile", (state) => MESSAGES.createdInstanceProfile .replace("${INSTANCE_PROFILE_NAME}", NAMES.instanceProfileName) .replace("${INSTANCE_PROFILE_ARN}", state.instanceProfileArn), ), new ScenarioOutput( "addingRoleToInstanceProfile", MESSAGES.addingRoleToInstanceProfile .replace("${INSTANCE_PROFILE_NAME}", NAMES.instanceProfileName) .replace("${INSTANCE_ROLE_NAME}", NAMES.instanceRoleName), ), new ScenarioAction("addRoleToInstanceProfile", () => { const client = new IAMClient({}); return client.send( new AddRoleToInstanceProfileCommand({ RoleName: NAMES.instanceRoleName, InstanceProfileName: NAMES.instanceProfileName, }), ); }), new ScenarioOutput( "addedRoleToInstanceProfile", MESSAGES.addedRoleToInstanceProfile .replace("${INSTANCE_PROFILE_NAME}", NAMES.instanceProfileName) .replace("${INSTANCE_ROLE_NAME}", NAMES.instanceRoleName), ), ...initParamsSteps, new ScenarioOutput("creatingLaunchTemplate", MESSAGES.creatingLaunchTemplate), new ScenarioAction("createLaunchTemplate", async () => { // snippet-start:[javascript.v3.wkflw.resilient.CreateLaunchTemplate] const ssmClient = new SSMClient({}); const { Parameter } = await ssmClient.send( new GetParameterCommand({ Name: "/aws/service/ami-amazon-linux-latest/amzn2-ami-hvm-x86_64-gp2", }), ); const ec2Client = new EC2Client({}); await ec2Client.send( new CreateLaunchTemplateCommand({ LaunchTemplateName: NAMES.launchTemplateName, LaunchTemplateData: { InstanceType: "t3.micro", ImageId: Parameter.Value, IamInstanceProfile: { Name: NAMES.instanceProfileName }, UserData: readFileSync( join(RESOURCES_PATH, "server_startup_script.sh"), ).toString("base64"), KeyName: NAMES.keyPairName, }, }), // snippet-end:[javascript.v3.wkflw.resilient.CreateLaunchTemplate] ); }), new ScenarioOutput( "createdLaunchTemplate", MESSAGES.createdLaunchTemplate.replace( "${LAUNCH_TEMPLATE_NAME}", NAMES.launchTemplateName, ), ), new ScenarioOutput( "creatingAutoScalingGroup", MESSAGES.creatingAutoScalingGroup.replace( "${AUTO_SCALING_GROUP_NAME}", NAMES.autoScalingGroupName, ), ), new ScenarioAction("createAutoScalingGroup", async (state) => { const ec2Client = new EC2Client({}); const { AvailabilityZones } = await ec2Client.send( new DescribeAvailabilityZonesCommand({}), ); state.availabilityZoneNames = AvailabilityZones.map((az) => az.ZoneName); const autoScalingClient = new AutoScalingClient({}); await retry({ intervalInMs: 1000, maxRetries: 30 }, () => autoScalingClient.send( new CreateAutoScalingGroupCommand({ AvailabilityZones: state.availabilityZoneNames, AutoScalingGroupName: NAMES.autoScalingGroupName, LaunchTemplate: { LaunchTemplateName: NAMES.launchTemplateName, Version: "$Default", }, MinSize: 3, MaxSize: 3, }), ), ); }), new ScenarioOutput( "createdAutoScalingGroup", /** * @param {{ availabilityZoneNames: string[] }} state */ (state) => MESSAGES.createdAutoScalingGroup .replace("${AUTO_SCALING_GROUP_NAME}", NAMES.autoScalingGroupName) .replace( "${AVAILABILITY_ZONE_NAMES}", state.availabilityZoneNames.join(", "), ), ), new ScenarioInput("confirmContinue", MESSAGES.confirmContinue, { type: "confirm", }), new ScenarioOutput("loadBalancer", MESSAGES.loadBalancer), new ScenarioOutput("gettingVpc", MESSAGES.gettingVpc), new ScenarioAction("getVpc", async (state) => { // snippet-start:[javascript.v3.wkflw.resilient.DescribeVpcs] const client = new EC2Client({}); const { Vpcs } = await client.send( new DescribeVpcsCommand({ Filters: [{ Name: "is-default", Values: ["true"] }], }), ); // snippet-end:[javascript.v3.wkflw.resilient.DescribeVpcs] state.defaultVpc = Vpcs[0].VpcId; }), new ScenarioOutput("gotVpc", (state) => MESSAGES.gotVpc.replace("${VPC_ID}", state.defaultVpc), ), new ScenarioOutput("gettingSubnets", MESSAGES.gettingSubnets), new ScenarioAction("getSubnets", async (state) => { // snippet-start:[javascript.v3.wkflw.resilient.DescribeSubnets] const client = new EC2Client({}); const { Subnets } = await client.send( new DescribeSubnetsCommand({ Filters: [ { Name: "vpc-id", Values: [state.defaultVpc] }, { Name: "availability-zone", Values: state.availabilityZoneNames }, { Name: "default-for-az", Values: ["true"] }, ], }), ); // snippet-end:[javascript.v3.wkflw.resilient.DescribeSubnets] state.subnets = Subnets.map((subnet) => subnet.SubnetId); }), new ScenarioOutput( "gotSubnets", /** * @param {{ subnets: string[] }} state */ (state) => MESSAGES.gotSubnets.replace("${SUBNETS}", state.subnets.join(", ")), ), new ScenarioOutput( "creatingLoadBalancerTargetGroup", MESSAGES.creatingLoadBalancerTargetGroup.replace( "${TARGET_GROUP_NAME}", NAMES.loadBalancerTargetGroupName, ), ), new ScenarioAction("createLoadBalancerTargetGroup", async (state) => { // snippet-start:[javascript.v3.wkflw.resilient.CreateTargetGroup] const client = new ElasticLoadBalancingV2Client({}); const { TargetGroups } = await client.send( new CreateTargetGroupCommand({ Name: NAMES.loadBalancerTargetGroupName, Protocol: "HTTP", Port: 80, HealthCheckPath: "/healthcheck", HealthCheckIntervalSeconds: 10, HealthCheckTimeoutSeconds: 5, HealthyThresholdCount: 2, UnhealthyThresholdCount: 2, VpcId: state.defaultVpc, }), ); // snippet-end:[javascript.v3.wkflw.resilient.CreateTargetGroup] const targetGroup = TargetGroups[0]; state.targetGroupArn = targetGroup.TargetGroupArn; state.targetGroupProtocol = targetGroup.Protocol; state.targetGroupPort = targetGroup.Port; }), new ScenarioOutput( "createdLoadBalancerTargetGroup", MESSAGES.createdLoadBalancerTargetGroup.replace( "${TARGET_GROUP_NAME}", NAMES.loadBalancerTargetGroupName, ), ), new ScenarioOutput( "creatingLoadBalancer", MESSAGES.creatingLoadBalancer.replace("${LB_NAME}", NAMES.loadBalancerName), ), new ScenarioAction("createLoadBalancer", async (state) => { // snippet-start:[javascript.v3.wkflw.resilient.CreateLoadBalancer] const client = new ElasticLoadBalancingV2Client({}); const { LoadBalancers } = await client.send( new CreateLoadBalancerCommand({ Name: NAMES.loadBalancerName, Subnets: state.subnets, }), ); state.loadBalancerDns = LoadBalancers[0].DNSName; state.loadBalancerArn = LoadBalancers[0].LoadBalancerArn; await waitUntilLoadBalancerAvailable( { client }, { Names: [NAMES.loadBalancerName] }, ); // snippet-end:[javascript.v3.wkflw.resilient.CreateLoadBalancer] }), new ScenarioOutput("createdLoadBalancer", (state) => MESSAGES.createdLoadBalancer .replace("${LB_NAME}", NAMES.loadBalancerName) .replace("${DNS_NAME}", state.loadBalancerDns), ), new ScenarioOutput( "creatingListener", MESSAGES.creatingLoadBalancerListener .replace("${LB_NAME}", NAMES.loadBalancerName) .replace("${TARGET_GROUP_NAME}", NAMES.loadBalancerTargetGroupName), ), new ScenarioAction("createListener", async (state) => { // snippet-start:[javascript.v3.wkflw.resilient.CreateListener] const client = new ElasticLoadBalancingV2Client({}); const { Listeners } = await client.send( new CreateListenerCommand({ LoadBalancerArn: state.loadBalancerArn, Protocol: state.targetGroupProtocol, Port: state.targetGroupPort, DefaultActions: [ { Type: "forward", TargetGroupArn: state.targetGroupArn }, ], }), ); // snippet-end:[javascript.v3.wkflw.resilient.CreateListener] const listener = Listeners[0]; state.loadBalancerListenerArn = listener.ListenerArn; }), new ScenarioOutput("createdListener", (state) => MESSAGES.createdLoadBalancerListener.replace( "${LB_LISTENER_ARN}", state.loadBalancerListenerArn, ), ), new ScenarioOutput( "attachingLoadBalancerTargetGroup", MESSAGES.attachingLoadBalancerTargetGroup .replace("${TARGET_GROUP_NAME}", NAMES.loadBalancerTargetGroupName) .replace("${AUTO_SCALING_GROUP_NAME}", NAMES.autoScalingGroupName), ), new ScenarioAction("attachLoadBalancerTargetGroup", async (state) => { // snippet-start:[javascript.v3.wkflw.resilient.AttachTargetGroup] const client = new AutoScalingClient({}); await client.send( new AttachLoadBalancerTargetGroupsCommand({ AutoScalingGroupName: NAMES.autoScalingGroupName, TargetGroupARNs: [state.targetGroupArn], }), ); // snippet-end:[javascript.v3.wkflw.resilient.AttachTargetGroup] }), new ScenarioOutput( "attachedLoadBalancerTargetGroup", MESSAGES.attachedLoadBalancerTargetGroup, ), new ScenarioOutput("verifyingInboundPort", MESSAGES.verifyingInboundPort), new ScenarioAction( "verifyInboundPort", /** * * @param {{ defaultSecurityGroup: import('@aws-sdk/client-ec2').SecurityGroup}} state */ async (state) => { const client = new EC2Client({}); const { SecurityGroups } = await client.send( new DescribeSecurityGroupsCommand({ Filters: [{ Name: "group-name", Values: ["default"] }], }), ); if (!SecurityGroups) { state.verifyInboundPortError = new Error(MESSAGES.noSecurityGroups); } state.defaultSecurityGroup = SecurityGroups[0]; /** * @type {string} */ const ipResponse = (await axios.get("http://checkip.amazonaws.com")).data; state.myIp = ipResponse.trim(); const myIpRules = state.defaultSecurityGroup.IpPermissions.filter( ({ IpRanges }) => IpRanges.some( ({ CidrIp }) => CidrIp.startsWith(state.myIp) || CidrIp === "0.0.0.0/0", ), ) .filter(({ IpProtocol }) => IpProtocol === "tcp") .filter(({ FromPort }) => FromPort === 80); state.myIpRules = myIpRules; }, ), new ScenarioOutput( "verifiedInboundPort", /** * @param {{ myIpRules: any[] }} state */ (state) => { if (state.myIpRules.length > 0) { return MESSAGES.foundIpRules.replace( "${IP_RULES}", JSON.stringify(state.myIpRules, null, 2), ); } else { return MESSAGES.noIpRules; } }, ), new ScenarioInput( "shouldAddInboundRule", /** * @param {{ myIpRules: any[] }} state */ (state) => { if (state.myIpRules.length > 0) { return false; } else { return MESSAGES.noIpRules; } }, { type: "confirm" }, ), new ScenarioAction( "addInboundRule", /** * @param {{ defaultSecurityGroup: import('@aws-sdk/client-ec2').SecurityGroup }} state */ async (state) => { if (!state.shouldAddInboundRule) { return; } const client = new EC2Client({}); await client.send( new AuthorizeSecurityGroupIngressCommand({ GroupId: state.defaultSecurityGroup.GroupId, CidrIp: `${state.myIp}/32`, FromPort: 80, ToPort: 80, IpProtocol: "tcp", }), ); }, ), new ScenarioOutput("addedInboundRule", (state) => { if (state.shouldAddInboundRule) { return MESSAGES.addedInboundRule.replace("${IP_ADDRESS}", state.myIp); } else { return false; } }), new ScenarioOutput("verifyingEndpoint", (state) => MESSAGES.verifyingEndpoint.replace("${DNS_NAME}", state.loadBalancerDns), ), new ScenarioAction("verifyEndpoint", async (state) => { try { const response = await retry({ intervalInMs: 2000, maxRetries: 30 }, () => axios.get(`http://${state.loadBalancerDns}`), ); state.endpointResponse = JSON.stringify(response.data, null, 2); } catch (e) { state.verifyEndpointError = e; } }), new ScenarioOutput("verifiedEndpoint", (state) => { if (state.verifyEndpointError) { console.error(state.verifyEndpointError); } else { return MESSAGES.verifiedEndpoint.replace( "${ENDPOINT_RESPONSE}", state.endpointResponse, ); } }), ];

Menyusun langkah-langkah untuk menjalankan demo.

// Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved. // SPDX-License-Identifier: Apache-2.0 import { readFileSync } from "node:fs"; import { join } from "node:path"; import axios from "axios"; import { DescribeTargetGroupsCommand, DescribeTargetHealthCommand, ElasticLoadBalancingV2Client, } from "@aws-sdk/client-elastic-load-balancing-v2"; import { DescribeInstanceInformationCommand, PutParameterCommand, SSMClient, SendCommandCommand, } from "@aws-sdk/client-ssm"; import { IAMClient, CreatePolicyCommand, CreateRoleCommand, AttachRolePolicyCommand, CreateInstanceProfileCommand, AddRoleToInstanceProfileCommand, waitUntilInstanceProfileExists, } from "@aws-sdk/client-iam"; import { AutoScalingClient, DescribeAutoScalingGroupsCommand, TerminateInstanceInAutoScalingGroupCommand, } from "@aws-sdk/client-auto-scaling"; import { DescribeIamInstanceProfileAssociationsCommand, EC2Client, RebootInstancesCommand, ReplaceIamInstanceProfileAssociationCommand, } from "@aws-sdk/client-ec2"; import { ScenarioAction, ScenarioInput, ScenarioOutput, } from "@aws-doc-sdk-examples/lib/scenario/scenario.js"; import { retry } from "@aws-doc-sdk-examples/lib/utils/util-timers.js"; import { MESSAGES, NAMES, RESOURCES_PATH } from "./constants.js"; import { findLoadBalancer } from "./shared.js"; const getRecommendation = new ScenarioAction( "getRecommendation", async (state) => { const loadBalancer = await findLoadBalancer(NAMES.loadBalancerName); if (loadBalancer) { state.loadBalancerDnsName = loadBalancer.DNSName; try { state.recommendation = ( await axios.get(`http://${state.loadBalancerDnsName}`) ).data; } catch (e) { state.recommendation = e instanceof Error ? e.message : e; } } else { throw new Error(MESSAGES.demoFindLoadBalancerError); } }, ); const getRecommendationResult = new ScenarioOutput( "getRecommendationResult", (state) => `Recommendation:\n${JSON.stringify(state.recommendation, null, 2)}`, { preformatted: true }, ); const getHealthCheck = new ScenarioAction("getHealthCheck", async (state) => { // snippet-start:[javascript.v3.wkflw.resilient.DescribeTargetGroups] const client = new ElasticLoadBalancingV2Client({}); const { TargetGroups } = await client.send( new DescribeTargetGroupsCommand({ Names: [NAMES.loadBalancerTargetGroupName], }), ); // snippet-end:[javascript.v3.wkflw.resilient.DescribeTargetGroups] // snippet-start:[javascript.v3.wkflw.resilient.DescribeTargetHealth] const { TargetHealthDescriptions } = await client.send( new DescribeTargetHealthCommand({ TargetGroupArn: TargetGroups[0].TargetGroupArn, }), ); // snippet-end:[javascript.v3.wkflw.resilient.DescribeTargetHealth] state.targetHealthDescriptions = TargetHealthDescriptions; }); const getHealthCheckResult = new ScenarioOutput( "getHealthCheckResult", /** * @param {{ targetHealthDescriptions: import('@aws-sdk/client-elastic-load-balancing-v2').TargetHealthDescription[]}} state */ (state) => { const status = state.targetHealthDescriptions .map((th) => `${th.Target.Id}: ${th.TargetHealth.State}`) .join("\n"); return `Health check:\n${status}`; }, { preformatted: true }, ); const loadBalancerLoop = new ScenarioAction( "loadBalancerLoop", getRecommendation.action, { whileConfig: { inputEquals: true, input: new ScenarioInput( "loadBalancerCheck", MESSAGES.demoLoadBalancerCheck, { type: "confirm", }, ), output: getRecommendationResult, }, }, ); const healthCheckLoop = new ScenarioAction( "healthCheckLoop", getHealthCheck.action, { whileConfig: { inputEquals: true, input: new ScenarioInput("healthCheck", MESSAGES.demoHealthCheck, { type: "confirm", }), output: getHealthCheckResult, }, }, ); const statusSteps = [ getRecommendation, getRecommendationResult, getHealthCheck, getHealthCheckResult, ]; /** * @type {import('@aws-doc-sdk-examples/lib/scenario.js').Step[]} */ export const demoSteps = [ new ScenarioOutput("header", MESSAGES.demoHeader, { header: true }), new ScenarioOutput("sanityCheck", MESSAGES.demoSanityCheck), ...statusSteps, new ScenarioInput( "brokenDependencyConfirmation", MESSAGES.demoBrokenDependencyConfirmation, { type: "confirm" }, ), new ScenarioAction("brokenDependency", async (state) => { if (!state.brokenDependencyConfirmation) { process.exit(); } else { const client = new SSMClient({}); state.badTableName = `fake-table-${Date.now()}`; await client.send( new PutParameterCommand({ Name: NAMES.ssmTableNameKey, Value: state.badTableName, Overwrite: true, Type: "String", }), ); } }), new ScenarioOutput("testBrokenDependency", (state) => MESSAGES.demoTestBrokenDependency.replace( "${TABLE_NAME}", state.badTableName, ), ), ...statusSteps, new ScenarioInput( "staticResponseConfirmation", MESSAGES.demoStaticResponseConfirmation, { type: "confirm" }, ), new ScenarioAction("staticResponse", async (state) => { if (!state.staticResponseConfirmation) { process.exit(); } else { const client = new SSMClient({}); await client.send( new PutParameterCommand({ Name: NAMES.ssmFailureResponseKey, Value: "static", Overwrite: true, Type: "String", }), ); } }), new ScenarioOutput("testStaticResponse", MESSAGES.demoTestStaticResponse), ...statusSteps, new ScenarioInput( "badCredentialsConfirmation", MESSAGES.demoBadCredentialsConfirmation, { type: "confirm" }, ), new ScenarioAction("badCredentialsExit", (state) => { if (!state.badCredentialsConfirmation) { process.exit(); } }), new ScenarioAction("fixDynamoDBName", async () => { const client = new SSMClient({}); await client.send( new PutParameterCommand({ Name: NAMES.ssmTableNameKey, Value: NAMES.tableName, Overwrite: true, Type: "String", }), ); }), new ScenarioAction( "badCredentials", /** * @param {{ targetInstance: import('@aws-sdk/client-auto-scaling').Instance }} state */ async (state) => { await createSsmOnlyInstanceProfile(); const autoScalingClient = new AutoScalingClient({}); const { AutoScalingGroups } = await autoScalingClient.send( new DescribeAutoScalingGroupsCommand({ AutoScalingGroupNames: [NAMES.autoScalingGroupName], }), ); state.targetInstance = AutoScalingGroups[0].Instances[0]; // snippet-start:[javascript.v3.wkflw.resilient.DescribeIamInstanceProfileAssociations] const ec2Client = new EC2Client({}); const { IamInstanceProfileAssociations } = await ec2Client.send( new DescribeIamInstanceProfileAssociationsCommand({ Filters: [ { Name: "instance-id", Values: [state.targetInstance.InstanceId] }, ], }), ); // snippet-end:[javascript.v3.wkflw.resilient.DescribeIamInstanceProfileAssociations] state.instanceProfileAssociationId = IamInstanceProfileAssociations[0].AssociationId; // snippet-start:[javascript.v3.wkflw.resilient.ReplaceIamInstanceProfileAssociation] await retry({ intervalInMs: 1000, maxRetries: 30 }, () => ec2Client.send( new ReplaceIamInstanceProfileAssociationCommand({ AssociationId: state.instanceProfileAssociationId, IamInstanceProfile: { Name: NAMES.ssmOnlyInstanceProfileName }, }), ), ); // snippet-end:[javascript.v3.wkflw.resilient.ReplaceIamInstanceProfileAssociation] await ec2Client.send( new RebootInstancesCommand({ InstanceIds: [state.targetInstance.InstanceId], }), ); const ssmClient = new SSMClient({}); await retry({ intervalInMs: 20000, maxRetries: 15 }, async () => { const { InstanceInformationList } = await ssmClient.send( new DescribeInstanceInformationCommand({}), ); const instance = InstanceInformationList.find( (info) => info.InstanceId === state.targetInstance.InstanceId, ); if (!instance) { throw new Error("Instance not found."); } }); await ssmClient.send( new SendCommandCommand({ InstanceIds: [state.targetInstance.InstanceId], DocumentName: "AWS-RunShellScript", Parameters: { commands: ["cd / && sudo python3 server.py 80"] }, }), ); }, ), new ScenarioOutput( "testBadCredentials", /** * @param {{ targetInstance: import('@aws-sdk/client-ssm').InstanceInformation}} state */ (state) => MESSAGES.demoTestBadCredentials.replace( "${INSTANCE_ID}", state.targetInstance.InstanceId, ), ), loadBalancerLoop, new ScenarioInput( "deepHealthCheckConfirmation", MESSAGES.demoDeepHealthCheckConfirmation, { type: "confirm" }, ), new ScenarioAction("deepHealthCheckExit", (state) => { if (!state.deepHealthCheckConfirmation) { process.exit(); } }), new ScenarioAction("deepHealthCheck", async () => { const client = new SSMClient({}); await client.send( new PutParameterCommand({ Name: NAMES.ssmHealthCheckKey, Value: "deep", Overwrite: true, Type: "String", }), ); }), new ScenarioOutput("testDeepHealthCheck", MESSAGES.demoTestDeepHealthCheck), healthCheckLoop, loadBalancerLoop, new ScenarioInput( "killInstanceConfirmation", /** * @param {{ targetInstance: import('@aws-sdk/client-ssm').InstanceInformation }} state */ (state) => MESSAGES.demoKillInstanceConfirmation.replace( "${INSTANCE_ID}", state.targetInstance.InstanceId, ), { type: "confirm" }, ), new ScenarioAction("killInstanceExit", (state) => { if (!state.killInstanceConfirmation) { process.exit(); } }), new ScenarioAction( "killInstance", /** * @param {{ targetInstance: import('@aws-sdk/client-ssm').InstanceInformation }} state */ async (state) => { const client = new AutoScalingClient({}); await client.send( new TerminateInstanceInAutoScalingGroupCommand({ InstanceId: state.targetInstance.InstanceId, ShouldDecrementDesiredCapacity: false, }), ); }, ), new ScenarioOutput("testKillInstance", MESSAGES.demoTestKillInstance), healthCheckLoop, loadBalancerLoop, new ScenarioInput("failOpenConfirmation", MESSAGES.demoFailOpenConfirmation, { type: "confirm", }), new ScenarioAction("failOpenExit", (state) => { if (!state.failOpenConfirmation) { process.exit(); } }), new ScenarioAction("failOpen", () => { const client = new SSMClient({}); return client.send( new PutParameterCommand({ Name: NAMES.ssmTableNameKey, Value: `fake-table-${Date.now()}`, Overwrite: true, Type: "String", }), ); }), new ScenarioOutput("testFailOpen", MESSAGES.demoFailOpenTest), healthCheckLoop, loadBalancerLoop, new ScenarioInput( "resetTableConfirmation", MESSAGES.demoResetTableConfirmation, { type: "confirm" }, ), new ScenarioAction("resetTableExit", (state) => { if (!state.resetTableConfirmation) { process.exit(); } }), new ScenarioAction("resetTable", async () => { const client = new SSMClient({}); await client.send( new PutParameterCommand({ Name: NAMES.ssmTableNameKey, Value: NAMES.tableName, Overwrite: true, Type: "String", }), ); }), new ScenarioOutput("testResetTable", MESSAGES.demoTestResetTable), healthCheckLoop, loadBalancerLoop, ]; async function createSsmOnlyInstanceProfile() { const iamClient = new IAMClient({}); const { Policy } = await iamClient.send( new CreatePolicyCommand({ PolicyName: NAMES.ssmOnlyPolicyName, PolicyDocument: readFileSync( join(RESOURCES_PATH, "ssm_only_policy.json"), ), }), ); await iamClient.send( new CreateRoleCommand({ RoleName: NAMES.ssmOnlyRoleName, AssumeRolePolicyDocument: JSON.stringify({ Version: "2012-10-17", Statement: [ { Effect: "Allow", Principal: { Service: "ec2.amazonaws.com" }, Action: "sts:AssumeRole", }, ], }), }), ); await iamClient.send( new AttachRolePolicyCommand({ RoleName: NAMES.ssmOnlyRoleName, PolicyArn: Policy.Arn, }), ); await iamClient.send( new AttachRolePolicyCommand({ RoleName: NAMES.ssmOnlyRoleName, PolicyArn: "arn:aws:iam::aws:policy/AmazonSSMManagedInstanceCore", }), ); // snippet-start:[javascript.v3.wkflw.resilient.CreateInstanceProfile] const { InstanceProfile } = await iamClient.send( new CreateInstanceProfileCommand({ InstanceProfileName: NAMES.ssmOnlyInstanceProfileName, }), ); await waitUntilInstanceProfileExists( { client: iamClient }, { InstanceProfileName: NAMES.ssmOnlyInstanceProfileName }, ); // snippet-end:[javascript.v3.wkflw.resilient.CreateInstanceProfile] await iamClient.send( new AddRoleToInstanceProfileCommand({ InstanceProfileName: NAMES.ssmOnlyInstanceProfileName, RoleName: NAMES.ssmOnlyRoleName, }), ); return InstanceProfile; }

Menyusun langkah-langkah untuk menghancurkan semua sumber daya.

// Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved. // SPDX-License-Identifier: Apache-2.0 import { unlinkSync } from "node:fs"; import { DynamoDBClient, DeleteTableCommand } from "@aws-sdk/client-dynamodb"; import { EC2Client, DeleteKeyPairCommand, DeleteLaunchTemplateCommand, } from "@aws-sdk/client-ec2"; import { IAMClient, DeleteInstanceProfileCommand, RemoveRoleFromInstanceProfileCommand, DeletePolicyCommand, DeleteRoleCommand, DetachRolePolicyCommand, paginateListPolicies, } from "@aws-sdk/client-iam"; import { AutoScalingClient, DeleteAutoScalingGroupCommand, TerminateInstanceInAutoScalingGroupCommand, UpdateAutoScalingGroupCommand, paginateDescribeAutoScalingGroups, } from "@aws-sdk/client-auto-scaling"; import { DeleteLoadBalancerCommand, DeleteTargetGroupCommand, DescribeTargetGroupsCommand, ElasticLoadBalancingV2Client, } from "@aws-sdk/client-elastic-load-balancing-v2"; import { ScenarioOutput, ScenarioInput, ScenarioAction, } from "@aws-doc-sdk-examples/lib/scenario/index.js"; import { retry } from "@aws-doc-sdk-examples/lib/utils/util-timers.js"; import { MESSAGES, NAMES } from "./constants.js"; import { findLoadBalancer } from "./shared.js"; /** * @type {import('@aws-doc-sdk-examples/lib/scenario.js').Step[]} */ export const destroySteps = [ new ScenarioInput("destroy", MESSAGES.destroy, { type: "confirm" }), new ScenarioAction( "abort", (state) => state.destroy === false && process.exit(), ), new ScenarioAction("deleteTable", async (c) => { try { const client = new DynamoDBClient({}); await client.send(new DeleteTableCommand({ TableName: NAMES.tableName })); } catch (e) { c.deleteTableError = e; } }), new ScenarioOutput("deleteTableResult", (state) => { if (state.deleteTableError) { console.error(state.deleteTableError); return MESSAGES.deleteTableError.replace( "${TABLE_NAME}", NAMES.tableName, ); } else { return MESSAGES.deletedTable.replace("${TABLE_NAME}", NAMES.tableName); } }), new ScenarioAction("deleteKeyPair", async (state) => { try { const client = new EC2Client({}); await client.send( new DeleteKeyPairCommand({ KeyName: NAMES.keyPairName }), ); unlinkSync(`${NAMES.keyPairName}.pem`); } catch (e) { state.deleteKeyPairError = e; } }), new ScenarioOutput("deleteKeyPairResult", (state) => { if (state.deleteKeyPairError) { console.error(state.deleteKeyPairError); return MESSAGES.deleteKeyPairError.replace( "${KEY_PAIR_NAME}", NAMES.keyPairName, ); } else { return MESSAGES.deletedKeyPair.replace( "${KEY_PAIR_NAME}", NAMES.keyPairName, ); } }), new ScenarioAction("detachPolicyFromRole", async (state) => { try { const client = new IAMClient({}); const policy = await findPolicy(NAMES.instancePolicyName); if (!policy) { state.detachPolicyFromRoleError = new Error( `Policy ${NAMES.instancePolicyName} not found.`, ); } else { await client.send( new DetachRolePolicyCommand({ RoleName: NAMES.instanceRoleName, PolicyArn: policy.Arn, }), ); } } catch (e) { state.detachPolicyFromRoleError = e; } }), new ScenarioOutput("detachedPolicyFromRole", (state) => { if (state.detachPolicyFromRoleError) { console.error(state.detachPolicyFromRoleError); return MESSAGES.detachPolicyFromRoleError .replace("${INSTANCE_POLICY_NAME}", NAMES.instancePolicyName) .replace("${INSTANCE_ROLE_NAME}", NAMES.instanceRoleName); } else { return MESSAGES.detachedPolicyFromRole .replace("${INSTANCE_POLICY_NAME}", NAMES.instancePolicyName) .replace("${INSTANCE_ROLE_NAME}", NAMES.instanceRoleName); } }), new ScenarioAction("deleteInstancePolicy", async (state) => { const client = new IAMClient({}); const policy = await findPolicy(NAMES.instancePolicyName); if (!policy) { state.deletePolicyError = new Error( `Policy ${NAMES.instancePolicyName} not found.`, ); } else { return client.send( new DeletePolicyCommand({ PolicyArn: policy.Arn, }), ); } }), new ScenarioOutput("deletePolicyResult", (state) => { if (state.deletePolicyError) { console.error(state.deletePolicyError); return MESSAGES.deletePolicyError.replace( "${INSTANCE_POLICY_NAME}", NAMES.instancePolicyName, ); } else { return MESSAGES.deletedPolicy.replace( "${INSTANCE_POLICY_NAME}", NAMES.instancePolicyName, ); } }), new ScenarioAction("removeRoleFromInstanceProfile", async (state) => { try { const client = new IAMClient({}); await client.send( new RemoveRoleFromInstanceProfileCommand({ RoleName: NAMES.instanceRoleName, InstanceProfileName: NAMES.instanceProfileName, }), ); } catch (e) { state.removeRoleFromInstanceProfileError = e; } }), new ScenarioOutput("removeRoleFromInstanceProfileResult", (state) => { if (state.removeRoleFromInstanceProfile) { console.error(state.removeRoleFromInstanceProfileError); return MESSAGES.removeRoleFromInstanceProfileError .replace("${INSTANCE_PROFILE_NAME}", NAMES.instanceProfileName) .replace("${INSTANCE_ROLE_NAME}", NAMES.instanceRoleName); } else { return MESSAGES.removedRoleFromInstanceProfile .replace("${INSTANCE_PROFILE_NAME}", NAMES.instanceProfileName) .replace("${INSTANCE_ROLE_NAME}", NAMES.instanceRoleName); } }), new ScenarioAction("deleteInstanceRole", async (state) => { try { const client = new IAMClient({}); await client.send( new DeleteRoleCommand({ RoleName: NAMES.instanceRoleName, }), ); } catch (e) { state.deleteInstanceRoleError = e; } }), new ScenarioOutput("deleteInstanceRoleResult", (state) => { if (state.deleteInstanceRoleError) { console.error(state.deleteInstanceRoleError); return MESSAGES.deleteInstanceRoleError.replace( "${INSTANCE_ROLE_NAME}", NAMES.instanceRoleName, ); } else { return MESSAGES.deletedInstanceRole.replace( "${INSTANCE_ROLE_NAME}", NAMES.instanceRoleName, ); } }), new ScenarioAction("deleteInstanceProfile", async (state) => { try { // snippet-start:[javascript.v3.wkflw.resilient.DeleteInstanceProfile] const client = new IAMClient({}); await client.send( new DeleteInstanceProfileCommand({ InstanceProfileName: NAMES.instanceProfileName, }), ); // snippet-end:[javascript.v3.wkflw.resilient.DeleteInstanceProfile] } catch (e) { state.deleteInstanceProfileError = e; } }), new ScenarioOutput("deleteInstanceProfileResult", (state) => { if (state.deleteInstanceProfileError) { console.error(state.deleteInstanceProfileError); return MESSAGES.deleteInstanceProfileError.replace( "${INSTANCE_PROFILE_NAME}", NAMES.instanceProfileName, ); } else { return MESSAGES.deletedInstanceProfile.replace( "${INSTANCE_PROFILE_NAME}", NAMES.instanceProfileName, ); } }), new ScenarioAction("deleteAutoScalingGroup", async (state) => { try { await terminateGroupInstances(NAMES.autoScalingGroupName); await retry({ intervalInMs: 60000, maxRetries: 60 }, async () => { await deleteAutoScalingGroup(NAMES.autoScalingGroupName); }); } catch (e) { state.deleteAutoScalingGroupError = e; } }), new ScenarioOutput("deleteAutoScalingGroupResult", (state) => { if (state.deleteAutoScalingGroupError) { console.error(state.deleteAutoScalingGroupError); return MESSAGES.deleteAutoScalingGroupError.replace( "${AUTO_SCALING_GROUP_NAME}", NAMES.autoScalingGroupName, ); } else { return MESSAGES.deletedAutoScalingGroup.replace( "${AUTO_SCALING_GROUP_NAME}", NAMES.autoScalingGroupName, ); } }), new ScenarioAction("deleteLaunchTemplate", async (state) => { const client = new EC2Client({}); try { // snippet-start:[javascript.v3.wkflw.resilient.DeleteLaunchTemplate] await client.send( new DeleteLaunchTemplateCommand({ LaunchTemplateName: NAMES.launchTemplateName, }), ); // snippet-end:[javascript.v3.wkflw.resilient.DeleteLaunchTemplate] } catch (e) { state.deleteLaunchTemplateError = e; } }), new ScenarioOutput("deleteLaunchTemplateResult", (state) => { if (state.deleteLaunchTemplateError) { console.error(state.deleteLaunchTemplateError); return MESSAGES.deleteLaunchTemplateError.replace( "${LAUNCH_TEMPLATE_NAME}", NAMES.launchTemplateName, ); } else { return MESSAGES.deletedLaunchTemplate.replace( "${LAUNCH_TEMPLATE_NAME}", NAMES.launchTemplateName, ); } }), new ScenarioAction("deleteLoadBalancer", async (state) => { try { // snippet-start:[javascript.v3.wkflw.resilient.DeleteLoadBalancer] const client = new ElasticLoadBalancingV2Client({}); const loadBalancer = await findLoadBalancer(NAMES.loadBalancerName); await client.send( new DeleteLoadBalancerCommand({ LoadBalancerArn: loadBalancer.LoadBalancerArn, }), ); await retry({ intervalInMs: 1000, maxRetries: 60 }, async () => { const lb = await findLoadBalancer(NAMES.loadBalancerName); if (lb) { throw new Error("Load balancer still exists."); } }); // snippet-end:[javascript.v3.wkflw.resilient.DeleteLoadBalancer] } catch (e) { state.deleteLoadBalancerError = e; } }), new ScenarioOutput("deleteLoadBalancerResult", (state) => { if (state.deleteLoadBalancerError) { console.error(state.deleteLoadBalancerError); return MESSAGES.deleteLoadBalancerError.replace( "${LB_NAME}", NAMES.loadBalancerName, ); } else { return MESSAGES.deletedLoadBalancer.replace( "${LB_NAME}", NAMES.loadBalancerName, ); } }), new ScenarioAction("deleteLoadBalancerTargetGroup", async (state) => { // snippet-start:[javascript.v3.wkflw.resilient.DeleteTargetGroup] const client = new ElasticLoadBalancingV2Client({}); try { const { TargetGroups } = await client.send( new DescribeTargetGroupsCommand({ Names: [NAMES.loadBalancerTargetGroupName], }), ); await retry({ intervalInMs: 1000, maxRetries: 30 }, () => client.send( new DeleteTargetGroupCommand({ TargetGroupArn: TargetGroups[0].TargetGroupArn, }), ), ); } catch (e) { state.deleteLoadBalancerTargetGroupError = e; } // snippet-end:[javascript.v3.wkflw.resilient.DeleteTargetGroup] }), new ScenarioOutput("deleteLoadBalancerTargetGroupResult", (state) => { if (state.deleteLoadBalancerTargetGroupError) { console.error(state.deleteLoadBalancerTargetGroupError); return MESSAGES.deleteLoadBalancerTargetGroupError.replace( "${TARGET_GROUP_NAME}", NAMES.loadBalancerTargetGroupName, ); } else { return MESSAGES.deletedLoadBalancerTargetGroup.replace( "${TARGET_GROUP_NAME}", NAMES.loadBalancerTargetGroupName, ); } }), new ScenarioAction("detachSsmOnlyRoleFromProfile", async (state) => { try { const client = new IAMClient({}); await client.send( new RemoveRoleFromInstanceProfileCommand({ InstanceProfileName: NAMES.ssmOnlyInstanceProfileName, RoleName: NAMES.ssmOnlyRoleName, }), ); } catch (e) { state.detachSsmOnlyRoleFromProfileError = e; } }), new ScenarioOutput("detachSsmOnlyRoleFromProfileResult", (state) => { if (state.detachSsmOnlyRoleFromProfileError) { console.error(state.detachSsmOnlyRoleFromProfileError); return MESSAGES.detachSsmOnlyRoleFromProfileError .replace("${ROLE_NAME}", NAMES.ssmOnlyRoleName) .replace("${PROFILE_NAME}", NAMES.ssmOnlyInstanceProfileName); } else { return MESSAGES.detachedSsmOnlyRoleFromProfile .replace("${ROLE_NAME}", NAMES.ssmOnlyRoleName) .replace("${PROFILE_NAME}", NAMES.ssmOnlyInstanceProfileName); } }), new ScenarioAction("detachSsmOnlyCustomRolePolicy", async (state) => { try { const iamClient = new IAMClient({}); const ssmOnlyPolicy = await findPolicy(NAMES.ssmOnlyPolicyName); await iamClient.send( new DetachRolePolicyCommand({ RoleName: NAMES.ssmOnlyRoleName, PolicyArn: ssmOnlyPolicy.Arn, }), ); } catch (e) { state.detachSsmOnlyCustomRolePolicyError = e; } }), new ScenarioOutput("detachSsmOnlyCustomRolePolicyResult", (state) => { if (state.detachSsmOnlyCustomRolePolicyError) { console.error(state.detachSsmOnlyCustomRolePolicyError); return MESSAGES.detachSsmOnlyCustomRolePolicyError .replace("${ROLE_NAME}", NAMES.ssmOnlyRoleName) .replace("${POLICY_NAME}", NAMES.ssmOnlyPolicyName); } else { return MESSAGES.detachedSsmOnlyCustomRolePolicy .replace("${ROLE_NAME}", NAMES.ssmOnlyRoleName) .replace("${POLICY_NAME}", NAMES.ssmOnlyPolicyName); } }), new ScenarioAction("detachSsmOnlyAWSRolePolicy", async (state) => { try { const iamClient = new IAMClient({}); await iamClient.send( new DetachRolePolicyCommand({ RoleName: NAMES.ssmOnlyRoleName, PolicyArn: "arn:aws:iam::aws:policy/AmazonSSMManagedInstanceCore", }), ); } catch (e) { state.detachSsmOnlyAWSRolePolicyError = e; } }), new ScenarioOutput("detachSsmOnlyAWSRolePolicyResult", (state) => { if (state.detachSsmOnlyAWSRolePolicyError) { console.error(state.detachSsmOnlyAWSRolePolicyError); return MESSAGES.detachSsmOnlyAWSRolePolicyError .replace("${ROLE_NAME}", NAMES.ssmOnlyRoleName) .replace("${POLICY_NAME}", "AmazonSSMManagedInstanceCore"); } else { return MESSAGES.detachedSsmOnlyAWSRolePolicy .replace("${ROLE_NAME}", NAMES.ssmOnlyRoleName) .replace("${POLICY_NAME}", "AmazonSSMManagedInstanceCore"); } }), new ScenarioAction("deleteSsmOnlyInstanceProfile", async (state) => { try { const iamClient = new IAMClient({}); await iamClient.send( new DeleteInstanceProfileCommand({ InstanceProfileName: NAMES.ssmOnlyInstanceProfileName, }), ); } catch (e) { state.deleteSsmOnlyInstanceProfileError = e; } }), new ScenarioOutput("deleteSsmOnlyInstanceProfileResult", (state) => { if (state.deleteSsmOnlyInstanceProfileError) { console.error(state.deleteSsmOnlyInstanceProfileError); return MESSAGES.deleteSsmOnlyInstanceProfileError.replace( "${INSTANCE_PROFILE_NAME}", NAMES.ssmOnlyInstanceProfileName, ); } else { return MESSAGES.deletedSsmOnlyInstanceProfile.replace( "${INSTANCE_PROFILE_NAME}", NAMES.ssmOnlyInstanceProfileName, ); } }), new ScenarioAction("deleteSsmOnlyPolicy", async (state) => { try { const iamClient = new IAMClient({}); const ssmOnlyPolicy = await findPolicy(NAMES.ssmOnlyPolicyName); await iamClient.send( new DeletePolicyCommand({ PolicyArn: ssmOnlyPolicy.Arn, }), ); } catch (e) { state.deleteSsmOnlyPolicyError = e; } }), new ScenarioOutput("deleteSsmOnlyPolicyResult", (state) => { if (state.deleteSsmOnlyPolicyError) { console.error(state.deleteSsmOnlyPolicyError); return MESSAGES.deleteSsmOnlyPolicyError.replace( "${POLICY_NAME}", NAMES.ssmOnlyPolicyName, ); } else { return MESSAGES.deletedSsmOnlyPolicy.replace( "${POLICY_NAME}", NAMES.ssmOnlyPolicyName, ); } }), new ScenarioAction("deleteSsmOnlyRole", async (state) => { try { const iamClient = new IAMClient({}); await iamClient.send( new DeleteRoleCommand({ RoleName: NAMES.ssmOnlyRoleName, }), ); } catch (e) { state.deleteSsmOnlyRoleError = e; } }), new ScenarioOutput("deleteSsmOnlyRoleResult", (state) => { if (state.deleteSsmOnlyRoleError) { console.error(state.deleteSsmOnlyRoleError); return MESSAGES.deleteSsmOnlyRoleError.replace( "${ROLE_NAME}", NAMES.ssmOnlyRoleName, ); } else { return MESSAGES.deletedSsmOnlyRole.replace( "${ROLE_NAME}", NAMES.ssmOnlyRoleName, ); } }), ]; /** * @param {string} policyName */ async function findPolicy(policyName) { const client = new IAMClient({}); const paginatedPolicies = paginateListPolicies({ client }, {}); for await (const page of paginatedPolicies) { const policy = page.Policies.find((p) => p.PolicyName === policyName); if (policy) { return policy; } } } /** * @param {string} groupName */ async function deleteAutoScalingGroup(groupName) { const client = new AutoScalingClient({}); try { await client.send( new DeleteAutoScalingGroupCommand({ AutoScalingGroupName: groupName, }), ); } catch (err) { if (!(err instanceof Error)) { throw err; } else { console.log(err.name); throw err; } } } /** * @param {string} groupName */ async function terminateGroupInstances(groupName) { const autoScalingClient = new AutoScalingClient({}); const group = await findAutoScalingGroup(groupName); await autoScalingClient.send( new UpdateAutoScalingGroupCommand({ AutoScalingGroupName: group.AutoScalingGroupName, MinSize: 0, }), ); for (const i of group.Instances) { await retry({ intervalInMs: 1000, maxRetries: 30 }, () => autoScalingClient.send( new TerminateInstanceInAutoScalingGroupCommand({ InstanceId: i.InstanceId, ShouldDecrementDesiredCapacity: true, }), ), ); } } async function findAutoScalingGroup(groupName) { const client = new AutoScalingClient({}); const paginatedGroups = paginateDescribeAutoScalingGroups({ client }, {}); for await (const page of paginatedGroups) { const group = page.AutoScalingGroups.find( (g) => g.AutoScalingGroupName === groupName, ); if (group) { return group; } } throw new Error(`Auto scaling group ${groupName} not found.`); }
Python
SDK for Python (Boto3)
catatan

Ada lebih banyak tentang GitHub. Temukan contoh lengkapnya dan pelajari cara pengaturan dan menjalankannya di Repositori Contoh Kode AWS.

Menjalankan skenario interaktif di prompt perintah.

class Runner: def __init__( self, resource_path, recommendation, autoscaler, loadbalancer, param_helper ): self.resource_path = resource_path self.recommendation = recommendation self.autoscaler = autoscaler self.loadbalancer = loadbalancer self.param_helper = param_helper self.protocol = "HTTP" self.port = 80 self.ssh_port = 22 def deploy(self): recommendations_path = f"{self.resource_path}/recommendations.json" startup_script = f"{self.resource_path}/server_startup_script.sh" instance_policy = f"{self.resource_path}/instance_policy.json" print( "\nFor this demo, we'll use the AWS SDK for Python (Boto3) to create several AWS resources\n" "to set up a load-balanced web service endpoint and explore some ways to make it resilient\n" "against various kinds of failures.\n\n" "Some of the resources create by this demo are:\n" ) print( "\t* A DynamoDB table that the web service depends on to provide book, movie, and song recommendations." ) print( "\t* An EC2 launch template that defines EC2 instances that each contain a Python web server." ) print( "\t* An EC2 Auto Scaling group that manages EC2 instances across several Availability Zones." ) print( "\t* An Elastic Load Balancing (ELB) load balancer that targets the Auto Scaling group to distribute requests." ) print("-" * 88) q.ask("Press Enter when you're ready to start deploying resources.") print( f"Creating and populating a DynamoDB table named '{self.recommendation.table_name}'." ) self.recommendation.create() self.recommendation.populate(recommendations_path) print("-" * 88) print( f"Creating an EC2 launch template that runs '{startup_script}' when an instance starts.\n" f"This script starts a Python web server defined in the `server.py` script. The web server\n" f"listens to HTTP requests on port 80 and responds to requests to '/' and to '/healthcheck'.\n" f"For demo purposes, this server is run as the root user. In production, the best practice is to\n" f"run a web server, such as Apache, with least-privileged credentials.\n" ) print( f"The template also defines an IAM policy that each instance uses to assume a role that grants\n" f"permissions to access the DynamoDB recommendation table and Systems Manager parameters\n" f"that control the flow of the demo.\n" ) self.autoscaler.create_template(startup_script, instance_policy) print("-" * 88) print( f"Creating an EC2 Auto Scaling group that maintains three EC2 instances, each in a different\n" f"Availability Zone." ) zones = self.autoscaler.create_group(3) print("-" * 88) print( "At this point, you have EC2 instances created. Once each instance starts, it listens for\n" "HTTP requests. You can see these instances in the console or continue with the demo." ) print("-" * 88) q.ask("Press Enter when you're ready to continue.") print(f"Creating variables that control the flow of the demo.\n") self.param_helper.reset() print( "\nCreating an Elastic Load Balancing target group and load balancer. The target group\n" "defines how the load balancer connects to instances. The load balancer provides a\n" "single endpoint where clients connect and dispatches requests to instances in the group.\n" ) vpc = self.autoscaler.get_default_vpc() subnets = self.autoscaler.get_subnets(vpc["VpcId"], zones) target_group = self.loadbalancer.create_target_group( self.protocol, self.port, vpc["VpcId"] ) self.loadbalancer.create_load_balancer( [subnet["SubnetId"] for subnet in subnets], target_group ) self.autoscaler.attach_load_balancer_target_group(target_group) print(f"Verifying access to the load balancer endpoint...") lb_success = self.loadbalancer.verify_load_balancer_endpoint() if not lb_success: print( "Couldn't connect to the load balancer, verifying that the port is open..." ) current_ip_address = requests.get( "http://checkip.amazonaws.com" ).text.strip() sec_group, port_is_open = self.autoscaler.verify_inbound_port( vpc, self.port, current_ip_address ) sec_group, ssh_port_is_open = self.autoscaler.verify_inbound_port( vpc, self.ssh_port, current_ip_address ) if not port_is_open: print( "For this example to work, the default security group for your default VPC must\n" "allows access from this computer. You can either add it automatically from this\n" "example or add it yourself using the AWS Management Console.\n" ) if q.ask( f"Do you want to add a rule to security group {sec_group['GroupId']} to allow\n" f"inbound traffic on port {self.port} from your computer's IP address of {current_ip_address}? (y/n) ", q.is_yesno, ): self.autoscaler.open_inbound_port( sec_group["GroupId"], self.port, current_ip_address ) if not ssh_port_is_open: if q.ask( f"Do you want to add a rule to security group {sec_group['GroupId']} to allow\n" f"inbound SSH traffic on port {self.ssh_port} for debugging from your computer's IP address of {current_ip_address}? (y/n) ", q.is_yesno, ): self.autoscaler.open_inbound_port( sec_group["GroupId"], self.ssh_port, current_ip_address ) lb_success = self.loadbalancer.verify_load_balancer_endpoint() if lb_success: print("Your load balancer is ready. You can access it by browsing to:\n") print(f"\thttp://{self.loadbalancer.endpoint()}\n") else: print( "Couldn't get a successful response from the load balancer endpoint. Troubleshoot by\n" "manually verifying that your VPC and security group are configured correctly and that\n" "you can successfully make a GET request to the load balancer endpoint:\n" ) print(f"\thttp://{self.loadbalancer.endpoint()}\n") print("-" * 88) q.ask("Press Enter when you're ready to continue with the demo.") def demo_choices(self): actions = [ "Send a GET request to the load balancer endpoint.", "Check the health of load balancer targets.", "Go to the next part of the demo.", ] choice = 0 while choice != 2: print("-" * 88) print( "\nSee the current state of the service by selecting one of the following choices:\n" ) choice = q.choose("\nWhich action would you like to take? ", actions) print("-" * 88) if choice == 0: print("Request:\n") print(f"GET http://{self.loadbalancer.endpoint()}") response = requests.get(f"http://{self.loadbalancer.endpoint()}") print("\nResponse:\n") print(f"{response.status_code}") if response.headers.get("content-type") == "application/json": pp(response.json()) elif choice == 1: print("\nChecking the health of load balancer targets:\n") health = self.loadbalancer.check_target_health() for target in health: state = target["TargetHealth"]["State"] print( f"\tTarget {target['Target']['Id']} on port {target['Target']['Port']} is {state}" ) if state != "healthy": print( f"\t\t{target['TargetHealth']['Reason']}: {target['TargetHealth']['Description']}\n" ) print( f"\nNote that it can take a minute or two for the health check to update\n" f"after changes are made.\n" ) elif choice == 2: print("\nOkay, let's move on.") print("-" * 88) def demo(self): ssm_only_policy = f"{self.resource_path}/ssm_only_policy.json" print("\nResetting parameters to starting values for demo.\n") self.param_helper.reset() print( "\nThis part of the demonstration shows how to toggle different parts of the system\n" "to create situations where the web service fails, and shows how using a resilient\n" "architecture can keep the web service running in spite of these failures." ) print("-" * 88) print( "At the start, the load balancer endpoint returns recommendations and reports that all targets are healthy." ) self.demo_choices() print( f"The web service running on the EC2 instances gets recommendations by querying a DynamoDB table.\n" f"The table name is contained in a Systems Manager parameter named '{self.param_helper.table}'.\n" f"To simulate a failure of the recommendation service, let's set this parameter to name a non-existent table.\n" ) self.param_helper.put(self.param_helper.table, "this-is-not-a-table") print( "\nNow, sending a GET request to the load balancer endpoint returns a failure code. But, the service reports as\n" "healthy to the load balancer because shallow health checks don't check for failure of the recommendation service." ) self.demo_choices() print( f"Instead of failing when the recommendation service fails, the web service can return a static response.\n" f"While this is not a perfect solution, it presents the customer with a somewhat better experience than failure.\n" ) self.param_helper.put(self.param_helper.failure_response, "static") print( f"\nNow, sending a GET request to the load balancer endpoint returns a static response.\n" f"The service still reports as healthy because health checks are still shallow.\n" ) self.demo_choices() print("Let's reinstate the recommendation service.\n") self.param_helper.put(self.param_helper.table, self.recommendation.table_name) print( "\nLet's also substitute bad credentials for one of the instances in the target group so that it can't\n" "access the DynamoDB recommendation table.\n" ) self.autoscaler.create_instance_profile( ssm_only_policy, self.autoscaler.bad_creds_policy_name, self.autoscaler.bad_creds_role_name, self.autoscaler.bad_creds_profile_name, ["AmazonSSMManagedInstanceCore"], ) instances = self.autoscaler.get_instances() bad_instance_id = instances[0] instance_profile = self.autoscaler.get_instance_profile(bad_instance_id) print( f"\nReplacing the profile for instance {bad_instance_id} with a profile that contains\n" f"bad credentials...\n" ) self.autoscaler.replace_instance_profile( bad_instance_id, self.autoscaler.bad_creds_profile_name, instance_profile["AssociationId"], ) print( "Now, sending a GET request to the load balancer endpoint returns either a recommendation or a static response,\n" "depending on which instance is selected by the load balancer.\n" ) self.demo_choices() print( "\nLet's implement a deep health check. For this demo, a deep health check tests whether\n" "the web service can access the DynamoDB table that it depends on for recommendations. Note that\n" "the deep health check is only for ELB routing and not for Auto Scaling instance health.\n" "This kind of deep health check is not recommended for Auto Scaling instance health, because it\n" "risks accidental termination of all instances in the Auto Scaling group when a dependent service fails.\n" ) print( "By implementing deep health checks, the load balancer can detect when one of the instances is failing\n" "and take that instance out of rotation.\n" ) self.param_helper.put(self.param_helper.health_check, "deep") print( f"\nNow, checking target health indicates that the instance with bad credentials ({bad_instance_id})\n" f"is unhealthy. Note that it might take a minute or two for the load balancer to detect the unhealthy \n" f"instance. Sending a GET request to the load balancer endpoint always returns a recommendation, because\n" "the load balancer takes unhealthy instances out of its rotation.\n" ) self.demo_choices() print( "\nBecause the instances in this demo are controlled by an auto scaler, the simplest way to fix an unhealthy\n" "instance is to terminate it and let the auto scaler start a new instance to replace it.\n" ) self.autoscaler.terminate_instance(bad_instance_id) print( "\nEven while the instance is terminating and the new instance is starting, sending a GET\n" "request to the web service continues to get a successful recommendation response because\n" "the load balancer routes requests to the healthy instances. After the replacement instance\n" "starts and reports as healthy, it is included in the load balancing rotation.\n" "\nNote that terminating and replacing an instance typically takes several minutes, during which time you\n" "can see the changing health check status until the new instance is running and healthy.\n" ) self.demo_choices() print( "\nIf the recommendation service fails now, deep health checks mean all instances report as unhealthy.\n" ) self.param_helper.put(self.param_helper.table, "this-is-not-a-table") print( "\nWhen all instances are unhealthy, the load balancer continues to route requests even to\n" "unhealthy instances, allowing them to fail open and return a static response rather than fail\n" "closed and report failure to the customer." ) self.demo_choices() self.param_helper.reset() def destroy(self): print( "This concludes the demo of how to build and manage a resilient service.\n" "To keep things tidy and to avoid unwanted charges on your account, we can clean up all AWS resources\n" "that were created for this demo." ) if q.ask("Do you want to clean up all demo resources? (y/n) ", q.is_yesno): self.loadbalancer.delete_load_balancer() self.loadbalancer.delete_target_group() self.autoscaler.delete_group() self.autoscaler.delete_key_pair() self.autoscaler.delete_template() self.autoscaler.delete_instance_profile( self.autoscaler.bad_creds_profile_name, self.autoscaler.bad_creds_role_name, ) self.recommendation.destroy() else: print( "Okay, we'll leave the resources intact.\n" "Don't forget to delete them when you're done with them or you might incur unexpected charges." ) def main(): parser = argparse.ArgumentParser() parser.add_argument( "--action", required=True, choices=["all", "deploy", "demo", "destroy"], help="The action to take for the demo. When 'all' is specified, resources are\n" "deployed, the demo is run, and resources are destroyed.", ) parser.add_argument( "--resource_path", default="../../../workflows/resilient_service/resources", help="The path to resource files used by this example, such as IAM policies and\n" "instance scripts.", ) args = parser.parse_args() print("-" * 88) print( "Welcome to the demonstration of How to Build and Manage a Resilient Service!" ) print("-" * 88) prefix = "doc-example-resilience" recommendation = RecommendationService.from_client( "doc-example-recommendation-service" ) autoscaler = AutoScaler.from_client(prefix) loadbalancer = LoadBalancer.from_client(prefix) param_helper = ParameterHelper.from_client(recommendation.table_name) runner = Runner( args.resource_path, recommendation, autoscaler, loadbalancer, param_helper ) actions = [args.action] if args.action != "all" else ["deploy", "demo", "destroy"] for action in actions: if action == "deploy": runner.deploy() elif action == "demo": runner.demo() elif action == "destroy": runner.destroy() print("-" * 88) print("Thanks for watching!") print("-" * 88) if __name__ == "__main__": logging.basicConfig(level=logging.INFO, format="%(levelname)s: %(message)s") main()

Membuat kelas yang menggabungkan tindakan Penskalaan Otomatis dan Amazon EC2.

class AutoScaler: """ Encapsulates Amazon EC2 Auto Scaling and EC2 management actions. """ def __init__( self, resource_prefix, inst_type, ami_param, autoscaling_client, ec2_client, ssm_client, iam_client, ): """ :param resource_prefix: The prefix for naming AWS resources that are created by this class. :param inst_type: The type of EC2 instance to create, such as t3.micro. :param ami_param: The Systems Manager parameter used to look up the AMI that is created. :param autoscaling_client: A Boto3 EC2 Auto Scaling client. :param ec2_client: A Boto3 EC2 client. :param ssm_client: A Boto3 Systems Manager client. :param iam_client: A Boto3 IAM client. """ self.inst_type = inst_type self.ami_param = ami_param self.autoscaling_client = autoscaling_client self.ec2_client = ec2_client self.ssm_client = ssm_client self.iam_client = iam_client self.launch_template_name = f"{resource_prefix}-template" self.group_name = f"{resource_prefix}-group" self.instance_policy_name = f"{resource_prefix}-pol" self.instance_role_name = f"{resource_prefix}-role" self.instance_profile_name = f"{resource_prefix}-prof" self.bad_creds_policy_name = f"{resource_prefix}-bc-pol" self.bad_creds_role_name = f"{resource_prefix}-bc-role" self.bad_creds_profile_name = f"{resource_prefix}-bc-prof" self.key_pair_name = f"{resource_prefix}-key-pair" @classmethod def from_client(cls, resource_prefix): """ Creates this class from Boto3 clients. :param resource_prefix: The prefix for naming AWS resources that are created by this class. """ as_client = boto3.client("autoscaling") ec2_client = boto3.client("ec2") ssm_client = boto3.client("ssm") iam_client = boto3.client("iam") return cls( resource_prefix, "t3.micro", "/aws/service/ami-amazon-linux-latest/amzn2-ami-hvm-x86_64-gp2", as_client, ec2_client, ssm_client, iam_client, ) def create_instance_profile( self, policy_file, policy_name, role_name, profile_name, aws_managed_policies=() ): """ Creates a policy, role, and profile that is associated with instances created by this class. An instance's associated profile defines a role that is assumed by the instance. The role has attached policies that specify the AWS permissions granted to clients that run on the instance. :param policy_file: The name of a JSON file that contains the policy definition to create and attach to the role. :param policy_name: The name to give the created policy. :param role_name: The name to give the created role. :param profile_name: The name to the created profile. :param aws_managed_policies: Additional AWS-managed policies that are attached to the role, such as AmazonSSMManagedInstanceCore to grant use of Systems Manager to send commands to the instance. :return: The ARN of the profile that is created. """ assume_role_doc = { "Version": "2012-10-17", "Statement": [ { "Effect": "Allow", "Principal": {"Service": "ec2.amazonaws.com"}, "Action": "sts:AssumeRole", } ], } with open(policy_file) as file: instance_policy_doc = file.read() policy_arn = None try: pol_response = self.iam_client.create_policy( PolicyName=policy_name, PolicyDocument=instance_policy_doc ) policy_arn = pol_response["Policy"]["Arn"] log.info("Created policy with ARN %s.", policy_arn) except ClientError as err: if err.response["Error"]["Code"] == "EntityAlreadyExists": log.info("Policy %s already exists, nothing to do.", policy_name) list_pol_response = self.iam_client.list_policies(Scope="Local") for pol in list_pol_response["Policies"]: if pol["PolicyName"] == policy_name: policy_arn = pol["Arn"] break if policy_arn is None: raise AutoScalerError(f"Couldn't create policy {policy_name}: {err}") try: self.iam_client.create_role( RoleName=role_name, AssumeRolePolicyDocument=json.dumps(assume_role_doc) ) self.iam_client.attach_role_policy(RoleName=role_name, PolicyArn=policy_arn) for aws_policy in aws_managed_policies: self.iam_client.attach_role_policy( RoleName=role_name, PolicyArn=f"arn:aws:iam::aws:policy/{aws_policy}", ) log.info("Created role %s and attached policy %s.", role_name, policy_arn) except ClientError as err: if err.response["Error"]["Code"] == "EntityAlreadyExists": log.info("Role %s already exists, nothing to do.", role_name) else: raise AutoScalerError(f"Couldn't create role {role_name}: {err}") try: profile_response = self.iam_client.create_instance_profile( InstanceProfileName=profile_name ) waiter = self.iam_client.get_waiter("instance_profile_exists") waiter.wait(InstanceProfileName=profile_name) time.sleep(10) # wait a little longer profile_arn = profile_response["InstanceProfile"]["Arn"] self.iam_client.add_role_to_instance_profile( InstanceProfileName=profile_name, RoleName=role_name ) log.info("Created profile %s and added role %s.", profile_name, role_name) except ClientError as err: if err.response["Error"]["Code"] == "EntityAlreadyExists": prof_response = self.iam_client.get_instance_profile( InstanceProfileName=profile_name ) profile_arn = prof_response["InstanceProfile"]["Arn"] log.info( "Instance profile %s already exists, nothing to do.", profile_name ) else: raise AutoScalerError( f"Couldn't create profile {profile_name} and attach it to role\n" f"{role_name}: {err}" ) return profile_arn def get_instance_profile(self, instance_id): """ Gets data about the profile associated with an instance. :param instance_id: The ID of the instance to look up. :return: The profile data. """ try: response = self.ec2_client.describe_iam_instance_profile_associations( Filters=[{"Name": "instance-id", "Values": [instance_id]}] ) except ClientError as err: raise AutoScalerError( f"Couldn't get instance profile association for instance {instance_id}: {err}" ) else: return response["IamInstanceProfileAssociations"][0] def replace_instance_profile( self, instance_id, new_instance_profile_name, profile_association_id ): """ Replaces the profile associated with a running instance. After the profile is replaced, the instance is rebooted to ensure that it uses the new profile. When the instance is ready, Systems Manager is used to restart the Python web server. :param instance_id: The ID of the instance to update. :param new_instance_profile_name: The name of the new profile to associate with the specified instance. :param profile_association_id: The ID of the existing profile association for the instance. """ try: self.ec2_client.replace_iam_instance_profile_association( IamInstanceProfile={"Name": new_instance_profile_name}, AssociationId=profile_association_id, ) log.info( "Replaced instance profile for association %s with profile %s.", profile_association_id, new_instance_profile_name, ) time.sleep(5) inst_ready = False tries = 0 while not inst_ready: if tries % 6 == 0: self.ec2_client.reboot_instances(InstanceIds=[instance_id]) log.info( "Rebooting instance %s and waiting for it to to be ready.", instance_id, ) tries += 1 time.sleep(10) response = self.ssm_client.describe_instance_information() for info in response["InstanceInformationList"]: if info["InstanceId"] == instance_id: inst_ready = True self.ssm_client.send_command( InstanceIds=[instance_id], DocumentName="AWS-RunShellScript", Parameters={"commands": ["cd / && sudo python3 server.py 80"]}, ) log.info("Restarted the Python web server on instance %s.", instance_id) except ClientError as err: raise AutoScalerError( f"Couldn't replace instance profile for association {profile_association_id}: {err}" ) def delete_instance_profile(self, profile_name, role_name): """ Detaches a role from an instance profile, detaches policies from the role, and deletes all the resources. :param profile_name: The name of the profile to delete. :param role_name: The name of the role to delete. """ try: self.iam_client.remove_role_from_instance_profile( InstanceProfileName=profile_name, RoleName=role_name ) self.iam_client.delete_instance_profile(InstanceProfileName=profile_name) log.info("Deleted instance profile %s.", profile_name) attached_policies = self.iam_client.list_attached_role_policies( RoleName=role_name ) for pol in attached_policies["AttachedPolicies"]: self.iam_client.detach_role_policy( RoleName=role_name, PolicyArn=pol["PolicyArn"] ) if not pol["PolicyArn"].startswith("arn:aws:iam::aws"): self.iam_client.delete_policy(PolicyArn=pol["PolicyArn"]) log.info("Detached and deleted policy %s.", pol["PolicyName"]) self.iam_client.delete_role(RoleName=role_name) log.info("Deleted role %s.", role_name) except ClientError as err: if err.response["Error"]["Code"] == "NoSuchEntity": log.info( "Instance profile %s doesn't exist, nothing to do.", profile_name ) else: raise AutoScalerError( f"Couldn't delete instance profile {profile_name} or detach " f"policies and delete role {role_name}: {err}" ) def create_key_pair(self, key_pair_name): """ Creates a new key pair. :param key_pair_name: The name of the key pair to create. :return: The newly created key pair. """ try: response = self.ec2_client.create_key_pair(KeyName=key_pair_name) with open(f"{key_pair_name}.pem", "w") as file: file.write(response["KeyMaterial"]) chmod(f"{key_pair_name}.pem", 0o600) log.info("Created key pair %s.", key_pair_name) except ClientError as err: raise AutoScalerError(f"Couldn't create key pair {key_pair_name}: {err}") def delete_key_pair(self): """ Deletes a key pair. :param key_pair_name: The name of the key pair to delete. """ try: self.ec2_client.delete_key_pair(KeyName=self.key_pair_name) remove(f"{self.key_pair_name}.pem") log.info("Deleted key pair %s.", self.key_pair_name) except ClientError as err: raise AutoScalerError( f"Couldn't delete key pair {self.key_pair_name}: {err}" ) except FileNotFoundError: log.info("Key pair %s doesn't exist, nothing to do.", self.key_pair_name) except PermissionError: log.info( "Inadequate permissions to delete key pair %s.", self.key_pair_name ) except Exception as err: raise AutoScalerError( f"Couldn't delete key pair {self.key_pair_name}: {err}" ) def create_template(self, server_startup_script_file, instance_policy_file): """ Creates an Amazon EC2 launch template to use with Amazon EC2 Auto Scaling. The launch template specifies a Bash script in its user data field that runs after the instance is started. This script installs Python packages and starts a Python web server on the instance. :param server_startup_script_file: The path to a Bash script file that is run when an instance starts. :param instance_policy_file: The path to a file that defines a permissions policy to create and attach to the instance profile. :return: Information about the newly created template. """ template = {} try: self.create_key_pair(self.key_pair_name) self.create_instance_profile( instance_policy_file, self.instance_policy_name, self.instance_role_name, self.instance_profile_name, ) with open(server_startup_script_file) as file: start_server_script = file.read() ami_latest = self.ssm_client.get_parameter(Name=self.ami_param) ami_id = ami_latest["Parameter"]["Value"] lt_response = self.ec2_client.create_launch_template( LaunchTemplateName=self.launch_template_name, LaunchTemplateData={ "InstanceType": self.inst_type, "ImageId": ami_id, "IamInstanceProfile": {"Name": self.instance_profile_name}, "UserData": base64.b64encode( start_server_script.encode(encoding="utf-8") ).decode(encoding="utf-8"), "KeyName": self.key_pair_name, }, ) template = lt_response["LaunchTemplate"] log.info( "Created launch template %s for AMI %s on %s.", self.launch_template_name, ami_id, self.inst_type, ) except ClientError as err: if ( err.response["Error"]["Code"] == "InvalidLaunchTemplateName.AlreadyExistsException" ): log.info( "Launch template %s already exists, nothing to do.", self.launch_template_name, ) else: raise AutoScalerError( f"Couldn't create launch template {self.launch_template_name}: {err}." ) return template def delete_template(self): """ Deletes a launch template. """ try: self.ec2_client.delete_launch_template( LaunchTemplateName=self.launch_template_name ) self.delete_instance_profile( self.instance_profile_name, self.instance_role_name ) log.info("Launch template %s deleted.", self.launch_template_name) except ClientError as err: if ( err.response["Error"]["Code"] == "InvalidLaunchTemplateName.NotFoundException" ): log.info( "Launch template %s does not exist, nothing to do.", self.launch_template_name, ) else: raise AutoScalerError( f"Couldn't delete launch template {self.launch_template_name}: {err}." ) def get_availability_zones(self): """ Gets a list of Availability Zones in the AWS Region of the Amazon EC2 client. :return: The list of Availability Zones for the client Region. """ try: response = self.ec2_client.describe_availability_zones() zones = [zone["ZoneName"] for zone in response["AvailabilityZones"]] except ClientError as err: raise AutoScalerError(f"Couldn't get availability zones: {err}.") else: return zones def create_group(self, group_size): """ Creates an EC2 Auto Scaling group with the specified size. :param group_size: The number of instances to set for the minimum and maximum in the group. :return: The list of Availability Zones specified for the group. """ zones = [] try: zones = self.get_availability_zones() self.autoscaling_client.create_auto_scaling_group( AutoScalingGroupName=self.group_name, AvailabilityZones=zones, LaunchTemplate={ "LaunchTemplateName": self.launch_template_name, "Version": "$Default", }, MinSize=group_size, MaxSize=group_size, ) log.info( "Created EC2 Auto Scaling group %s with availability zones %s.", self.launch_template_name, zones, ) except ClientError as err: if err.response["Error"]["Code"] == "AlreadyExists": log.info( "EC2 Auto Scaling group %s already exists, nothing to do.", self.group_name, ) else: raise AutoScalerError( f"Couldn't create EC2 Auto Scaling group {self.group_name}: {err}" ) return zones def get_instances(self): """ Gets data about the instances in the EC2 Auto Scaling group. :return: Data about the instances. """ try: as_response = self.autoscaling_client.describe_auto_scaling_groups( AutoScalingGroupNames=[self.group_name] ) instance_ids = [ i["InstanceId"] for i in as_response["AutoScalingGroups"][0]["Instances"] ] except ClientError as err: raise AutoScalerError( f"Couldn't get instances for Auto Scaling group {self.group_name}: {err}" ) else: return instance_ids def terminate_instance(self, instance_id): """ Terminates and instances in an EC2 Auto Scaling group. After an instance is terminated, it can no longer be accessed. :param instance_id: The ID of the instance to terminate. """ try: self.autoscaling_client.terminate_instance_in_auto_scaling_group( InstanceId=instance_id, ShouldDecrementDesiredCapacity=False ) log.info("Terminated instance %s.", instance_id) except ClientError as err: raise AutoScalerError(f"Couldn't terminate instance {instance_id}: {err}") def attach_load_balancer_target_group(self, lb_target_group): """ Attaches an Elastic Load Balancing (ELB) target group to this EC2 Auto Scaling group. The target group specifies how the load balancer forward requests to the instances in the group. :param lb_target_group: Data about the ELB target group to attach. """ try: self.autoscaling_client.attach_load_balancer_target_groups( AutoScalingGroupName=self.group_name, TargetGroupARNs=[lb_target_group["TargetGroupArn"]], ) log.info( "Attached load balancer target group %s to auto scaling group %s.", lb_target_group["TargetGroupName"], self.group_name, ) except ClientError as err: raise AutoScalerError( f"Couldn't attach load balancer target group {lb_target_group['TargetGroupName']}\n" f"to auto scaling group {self.group_name}" ) def _try_terminate_instance(self, inst_id): stopping = False log.info(f"Stopping {inst_id}.") while not stopping: try: self.autoscaling_client.terminate_instance_in_auto_scaling_group( InstanceId=inst_id, ShouldDecrementDesiredCapacity=True ) stopping = True except ClientError as err: if err.response["Error"]["Code"] == "ScalingActivityInProgress": log.info("Scaling activity in progress for %s. Waiting...", inst_id) time.sleep(10) else: raise AutoScalerError(f"Couldn't stop instance {inst_id}: {err}.") def _try_delete_group(self): """ Tries to delete the EC2 Auto Scaling group. If the group is in use or in progress, the function waits and retries until the group is successfully deleted. """ stopped = False while not stopped: try: self.autoscaling_client.delete_auto_scaling_group( AutoScalingGroupName=self.group_name ) stopped = True log.info("Deleted EC2 Auto Scaling group %s.", self.group_name) except ClientError as err: if ( err.response["Error"]["Code"] == "ResourceInUse" or err.response["Error"]["Code"] == "ScalingActivityInProgress" ): log.info( "Some instances are still running. Waiting for them to stop..." ) time.sleep(10) else: raise AutoScalerError( f"Couldn't delete group {self.group_name}: {err}." ) def delete_group(self): """ Terminates all instances in the group, deletes the EC2 Auto Scaling group. """ try: response = self.autoscaling_client.describe_auto_scaling_groups( AutoScalingGroupNames=[self.group_name] ) groups = response.get("AutoScalingGroups", []) if len(groups) > 0: self.autoscaling_client.update_auto_scaling_group( AutoScalingGroupName=self.group_name, MinSize=0 ) instance_ids = [inst["InstanceId"] for inst in groups[0]["Instances"]] for inst_id in instance_ids: self._try_terminate_instance(inst_id) self._try_delete_group() else: log.info("No groups found named %s, nothing to do.", self.group_name) except ClientError as err: raise AutoScalerError(f"Couldn't delete group {self.group_name}: {err}.") def get_default_vpc(self): """ Gets the default VPC for the account. :return: Data about the default VPC. """ try: response = self.ec2_client.describe_vpcs( Filters=[{"Name": "is-default", "Values": ["true"]}] ) except ClientError as err: raise AutoScalerError(f"Couldn't get default VPC: {err}") else: return response["Vpcs"][0] def verify_inbound_port(self, vpc, port, ip_address): """ Verify the default security group of the specified VPC allows ingress from this computer. This can be done by allowing ingress from this computer's IP address. In some situations, such as connecting from a corporate network, you must instead specify a prefix list ID. You can also temporarily open the port to any IP address while running this example. If you do, be sure to remove public access when you're done. :param vpc: The VPC used by this example. :param port: The port to verify. :param ip_address: This computer's IP address. :return: The default security group of the specific VPC, and a value that indicates whether the specified port is open. """ try: response = self.ec2_client.describe_security_groups( Filters=[ {"Name": "group-name", "Values": ["default"]}, {"Name": "vpc-id", "Values": [vpc["VpcId"]]}, ] ) sec_group = response["SecurityGroups"][0] port_is_open = False log.info("Found default security group %s.", sec_group["GroupId"]) for ip_perm in sec_group["IpPermissions"]: if ip_perm.get("FromPort", 0) == port: log.info("Found inbound rule: %s", ip_perm) for ip_range in ip_perm["IpRanges"]: cidr = ip_range.get("CidrIp", "") if cidr.startswith(ip_address) or cidr == "0.0.0.0/0": port_is_open = True if ip_perm["PrefixListIds"]: port_is_open = True if not port_is_open: log.info( "The inbound rule does not appear to be open to either this computer's IP\n" "address of %s, to all IP addresses (0.0.0.0/0), or to a prefix list ID.", ip_address, ) else: break except ClientError as err: raise AutoScalerError( f"Couldn't verify inbound rule for port {port} for VPC {vpc['VpcId']}: {err}" ) else: return sec_group, port_is_open def open_inbound_port(self, sec_group_id, port, ip_address): """ Add an ingress rule to the specified security group that allows access on the specified port from the specified IP address. :param sec_group_id: The ID of the security group to modify. :param port: The port to open. :param ip_address: The IP address that is granted access. """ try: self.ec2_client.authorize_security_group_ingress( GroupId=sec_group_id, CidrIp=f"{ip_address}/32", FromPort=port, ToPort=port, IpProtocol="tcp", ) log.info( "Authorized ingress to %s on port %s from %s.", sec_group_id, port, ip_address, ) except ClientError as err: raise AutoScalerError( f"Couldn't authorize ingress to {sec_group_id} on port {port} from {ip_address}: {err}" ) def get_subnets(self, vpc_id, zones): """ Gets the default subnets in a VPC for a specified list of Availability Zones. :param vpc_id: The ID of the VPC to look up. :param zones: The list of Availability Zones to look up. :return: The list of subnets found. """ try: response = self.ec2_client.describe_subnets( Filters=[ {"Name": "vpc-id", "Values": [vpc_id]}, {"Name": "availability-zone", "Values": zones}, {"Name": "default-for-az", "Values": ["true"]}, ] ) subnets = response["Subnets"] log.info("Found %s subnets for the specified zones.", len(subnets)) except ClientError as err: raise AutoScalerError(f"Couldn't get subnets: {err}") else: return subnets

Membuat kelas yang menggabungkan tindakan Penyeimbangan Beban Elastis.

class LoadBalancer: """Encapsulates Elastic Load Balancing (ELB) actions.""" def __init__(self, target_group_name, load_balancer_name, elb_client): """ :param target_group_name: The name of the target group associated with the load balancer. :param load_balancer_name: The name of the load balancer. :param elb_client: A Boto3 Elastic Load Balancing client. """ self.target_group_name = target_group_name self.load_balancer_name = load_balancer_name self.elb_client = elb_client self._endpoint = None @classmethod def from_client(cls, resource_prefix): """ Creates this class from a Boto3 client. :param resource_prefix: The prefix to give to AWS resources created by this class. """ elb_client = boto3.client("elbv2") return cls(f"{resource_prefix}-tg", f"{resource_prefix}-lb", elb_client) def endpoint(self): """ Gets the HTTP endpoint of the load balancer. :return: The endpoint. """ if self._endpoint is None: try: response = self.elb_client.describe_load_balancers( Names=[self.load_balancer_name] ) self._endpoint = response["LoadBalancers"][0]["DNSName"] except ClientError as err: raise LoadBalancerError( f"Couldn't get the endpoint for load balancer {self.load_balancer_name}: {err}" ) return self._endpoint def create_target_group(self, protocol, port, vpc_id): """ Creates an Elastic Load Balancing target group. The target group specifies how the load balancer forward requests to instances in the group and how instance health is checked. To speed up this demo, the health check is configured with shortened times and lower thresholds. In production, you might want to decrease the sensitivity of your health checks to avoid unwanted failures. :param protocol: The protocol to use to forward requests, such as 'HTTP'. :param port: The port to use to forward requests, such as 80. :param vpc_id: The ID of the VPC in which the load balancer exists. :return: Data about the newly created target group. """ try: response = self.elb_client.create_target_group( Name=self.target_group_name, Protocol=protocol, Port=port, HealthCheckPath="/healthcheck", HealthCheckIntervalSeconds=10, HealthCheckTimeoutSeconds=5, HealthyThresholdCount=2, UnhealthyThresholdCount=2, VpcId=vpc_id, ) target_group = response["TargetGroups"][0] log.info("Created load balancing target group %s.", self.target_group_name) except ClientError as err: raise LoadBalancerError( f"Couldn't create load balancing target group {self.target_group_name}: {err}" ) else: return target_group def delete_target_group(self): """ Deletes the target group. """ done = False while not done: try: response = self.elb_client.describe_target_groups( Names=[self.target_group_name] ) tg_arn = response["TargetGroups"][0]["TargetGroupArn"] self.elb_client.delete_target_group(TargetGroupArn=tg_arn) log.info( "Deleted load balancing target group %s.", self.target_group_name ) done = True except ClientError as err: if err.response["Error"]["Code"] == "TargetGroupNotFound": log.info( "Load balancer target group %s not found, nothing to do.", self.target_group_name, ) done = True elif err.response["Error"]["Code"] == "ResourceInUse": log.info( "Target group not yet released from load balancer, waiting..." ) time.sleep(10) else: raise LoadBalancerError( f"Couldn't delete load balancing target group {self.target_group_name}: {err}" ) def create_load_balancer(self, subnet_ids, target_group): """ Creates an Elastic Load Balancing load balancer that uses the specified subnets and forwards requests to the specified target group. :param subnet_ids: A list of subnets to associate with the load balancer. :param target_group: An existing target group that is added as a listener to the load balancer. :return: Data about the newly created load balancer. """ try: response = self.elb_client.create_load_balancer( Name=self.load_balancer_name, Subnets=subnet_ids ) load_balancer = response["LoadBalancers"][0] log.info("Created load balancer %s.", self.load_balancer_name) waiter = self.elb_client.get_waiter("load_balancer_available") log.info("Waiting for load balancer to be available...") waiter.wait(Names=[self.load_balancer_name]) log.info("Load balancer is available!") self.elb_client.create_listener( LoadBalancerArn=load_balancer["LoadBalancerArn"], Protocol=target_group["Protocol"], Port=target_group["Port"], DefaultActions=[ { "Type": "forward", "TargetGroupArn": target_group["TargetGroupArn"], } ], ) log.info( "Created listener to forward traffic from load balancer %s to target group %s.", self.load_balancer_name, target_group["TargetGroupName"], ) except ClientError as err: raise LoadBalancerError( f"Failed to create load balancer {self.load_balancer_name}" f"and add a listener for target group {target_group['TargetGroupName']}: {err}" ) else: self._endpoint = load_balancer["DNSName"] return load_balancer def delete_load_balancer(self): """ Deletes a load balancer. """ try: response = self.elb_client.describe_load_balancers( Names=[self.load_balancer_name] ) lb_arn = response["LoadBalancers"][0]["LoadBalancerArn"] self.elb_client.delete_load_balancer(LoadBalancerArn=lb_arn) log.info("Deleted load balancer %s.", self.load_balancer_name) waiter = self.elb_client.get_waiter("load_balancers_deleted") log.info("Waiting for load balancer to be deleted...") waiter.wait(Names=[self.load_balancer_name]) except ClientError as err: if err.response["Error"]["Code"] == "LoadBalancerNotFound": log.info( "Load balancer %s does not exist, nothing to do.", self.load_balancer_name, ) else: raise LoadBalancerError( f"Couldn't delete load balancer {self.load_balancer_name}: {err}" ) def verify_load_balancer_endpoint(self): """ Verify this computer can successfully send a GET request to the load balancer endpoint. """ success = False retries = 3 while not success and retries > 0: try: lb_response = requests.get(f"http://{self.endpoint()}") log.info( "Got response %s from load balancer endpoint.", lb_response.status_code, ) if lb_response.status_code == 200: success = True else: retries = 0 except requests.exceptions.ConnectionError: log.info( "Got connection error from load balancer endpoint, retrying..." ) retries -= 1 time.sleep(10) return success def check_target_health(self): """ Checks the health of the instances in the target group. :return: The health status of the target group. """ try: tg_response = self.elb_client.describe_target_groups( Names=[self.target_group_name] ) health_response = self.elb_client.describe_target_health( TargetGroupArn=tg_response["TargetGroups"][0]["TargetGroupArn"] ) except ClientError as err: raise LoadBalancerError( f"Couldn't check health of {self.target_group_name} targets: {err}" ) else: return health_response["TargetHealthDescriptions"]

Membuat kelas yang menggunakan DynamoDB untuk menyimulasikan layanan yang direkomendasikan.

class RecommendationService: """ Encapsulates a DynamoDB table to use as a service that recommends books, movies, and songs. """ def __init__(self, table_name, dynamodb_client): """ :param table_name: The name of the DynamoDB recommendations table. :param dynamodb_client: A Boto3 DynamoDB client. """ self.table_name = table_name self.dynamodb_client = dynamodb_client @classmethod def from_client(cls, table_name): """ Creates this class from a Boto3 client. :param table_name: The name of the DynamoDB recommendations table. """ ddb_client = boto3.client("dynamodb") return cls(table_name, ddb_client) def create(self): """ Creates a DynamoDB table to use a recommendation service. The table has a hash key named 'MediaType' that defines the type of media recommended, such as Book or Movie, and a range key named 'ItemId' that, combined with the MediaType, forms a unique identifier for the recommended item. :return: Data about the newly created table. """ try: response = self.dynamodb_client.create_table( TableName=self.table_name, AttributeDefinitions=[ {"AttributeName": "MediaType", "AttributeType": "S"}, {"AttributeName": "ItemId", "AttributeType": "N"}, ], KeySchema=[ {"AttributeName": "MediaType", "KeyType": "HASH"}, {"AttributeName": "ItemId", "KeyType": "RANGE"}, ], ProvisionedThroughput={"ReadCapacityUnits": 5, "WriteCapacityUnits": 5}, ) log.info("Creating table %s...", self.table_name) waiter = self.dynamodb_client.get_waiter("table_exists") waiter.wait(TableName=self.table_name) log.info("Table %s created.", self.table_name) except ClientError as err: if err.response["Error"]["Code"] == "ResourceInUseException": log.info("Table %s exists, nothing to be do.", self.table_name) else: raise RecommendationServiceError( self.table_name, f"ClientError when creating table: {err}." ) else: return response def populate(self, data_file): """ Populates the recommendations table from a JSON file. :param data_file: The path to the data file. """ try: with open(data_file) as data: items = json.load(data) batch = [{"PutRequest": {"Item": item}} for item in items] self.dynamodb_client.batch_write_item(RequestItems={self.table_name: batch}) log.info( "Populated table %s with items from %s.", self.table_name, data_file ) except ClientError as err: raise RecommendationServiceError( self.table_name, f"Couldn't populate table from {data_file}: {err}" ) def destroy(self): """ Deletes the recommendations table. """ try: self.dynamodb_client.delete_table(TableName=self.table_name) log.info("Deleting table %s...", self.table_name) waiter = self.dynamodb_client.get_waiter("table_not_exists") waiter.wait(TableName=self.table_name) log.info("Table %s deleted.", self.table_name) except ClientError as err: if err.response["Error"]["Code"] == "ResourceNotFoundException": log.info("Table %s does not exist, nothing to do.", self.table_name) else: raise RecommendationServiceError( self.table_name, f"ClientError when deleting table: {err}." )

Membuat kelas yang mengabungkan tindakan Systems Manager.

class ParameterHelper: """ Encapsulates Systems Manager parameters. This example uses these parameters to drive the demonstration of resilient architecture, such as failure of a dependency or how the service responds to a health check. """ table = "doc-example-resilient-architecture-table" failure_response = "doc-example-resilient-architecture-failure-response" health_check = "doc-example-resilient-architecture-health-check" def __init__(self, table_name, ssm_client): """ :param table_name: The name of the DynamoDB table that is used as a recommendation service. :param ssm_client: A Boto3 Systems Manager client. """ self.ssm_client = ssm_client self.table_name = table_name @classmethod def from_client(cls, table_name): ssm_client = boto3.client("ssm") return cls(table_name, ssm_client) def reset(self): """ Resets the Systems Manager parameters to starting values for the demo. These are the name of the DynamoDB recommendation table, no response when a dependency fails, and shallow health checks. """ self.put(self.table, self.table_name) self.put(self.failure_response, "none") self.put(self.health_check, "shallow") def put(self, name, value): """ Sets the value of a named Systems Manager parameter. :param name: The name of the parameter. :param value: The new value of the parameter. """ try: self.ssm_client.put_parameter( Name=name, Value=value, Overwrite=True, Type="String" ) log.info("Setting demo parameter %s to '%s'.", name, value) except ClientError as err: raise ParameterHelperError( f"Couldn't set parameter {name} to {value}: {err}" )

Untuk daftar lengkap panduan pengembang AWS SDK dan contoh kode, lihatMenggunakan layanan ini dengan AWS SDK. Topik ini juga mencakup informasi tentang cara memulai dan detail versi-versi SDK sebelumnya.