Class Stack

A root construct which represents a single CloudFormation stack.

Synopsis

Constructors

constructor(scope, name, props)

Creates a new stack.

Properties

account	The AWS account into which this stack will be deployed.
availabilityZones	Returnst the list of AZs that are availability in the AWS environment (account/region) associated with this stack.
dependencies	Return the stacks this stack depends on
environment	The environment coordinates in which this stack is deployed. In the form aws://account/region. Use stack.account and stack.region to obtain the specific values, no need to parse.
nested	Indicates if this is a nested stack, in which case parentStack will include a reference to it's parent.
notificationArns	Returns the list of notification Amazon Resource Names (ARNs) for the current stack.
parentStack	Returns the parent stack if this stack is nested.
partition	The partition in which this stack is defined
region	The AWS region into which this stack will be deployed (e.g. us-west-2).
stackId	The ID of the stack
stackName	The concrete CloudFormation physical stack name.
tags	Tags to be applied to the stack.
templateFile	The name of the CloudFormation template file emitted to the output directory during synthesis.
templateOptions	Options for CloudFormation template (like version, transform, description).
templateUrl	An attribute (late-bound) that represents the URL of the template file in the deployment bucket.
urlSuffix	The Amazon domain suffix for the region in which this stack is defined

Methods

addDependency(stack, reason)	Add a dependency between this stack and another stack
addDockerImageAsset(asset)
addFileAsset(asset)
allocateLogicalId(cfnElement)	Returns the naming scheme used to allocate logical IDs. By default, uses the HashedAddressingScheme but this method can be overridden to customize this behavior.
formatArn(components)	Creates an ARN from components.
getLogicalId(element)	Allocates a stack-unique CloudFormation-compatible logical identity for a specific resource.
isStack(x)	Return whether the given object is a Stack.
of(construct)	Looks up the first stack scope in which construct is defined. Fails if there is no stack up the tree.
parseArn(arn, sepIfToken, hasName)	Given an ARN, parses it and returns components.
prepare()	Prepare stack
prepareCrossReference(sourceStack, reference)	Exports a resolvable value for use in another stack.
renameLogicalId(oldId, newId)	Rename a generated logical identities
reportMissingContext(report)	Indicate that a context key was expected
resolve(obj)	Resolve a tokenized value in the context of the current stack.
synthesize(session)
toJsonString(obj, space)	Convert an object, potentially containing tokens, to a JSON string

Constructors

constructor(scope, name, props)

Creates a new stack.

Declaration

constructor(scope?: Construct, name?: string, props?: StackProps);

Parameters

scope Construct

name string

props StackProps

Properties

account

The AWS account into which this stack will be deployed.

This value is resolved according to the following rules:

1. The value provided to env.account when the stack is defined. This can either be a concerete account (e.g. 585695031111) or the Aws.accountId token. 3. Aws.accountId, which represents the CloudFormation intrinsic reference { "Ref": "AWS::AccountId" } encoded as a string token.

Preferably, you should use the return value as an opaque string and not attempt to parse it to implement your logic. If you do, you must first check that it is a concerete value an not an unresolved token. If this value is an unresolved token (Token.isUnresolved(stack.account) returns true), this implies that the user wishes that this stack will synthesize into a **account-agnostic template**. In this case, your code should either fail (throw an error, emit a synth error using node.addError) or implement some other region-agnostic behavior.

Declaration

readonly account: string;

Property Value

string

availabilityZones

Returnst the list of AZs that are availability in the AWS environment (account/region) associated with this stack.

If the stack is environment-agnostic (either account and/or region are tokens), this property will return an array with 2 tokens that will resolve at deploy-time to the first two availability zones returned from CloudFormation's Fn::GetAZs intrinsic function.

If they are not available in the context, returns a set of dummy values and reports them as missing, and let the CLI resolve them by calling EC2 DescribeAvailabilityZones on the target environment.

Declaration

readonly availabilityZones: string[];

Property Value

string[]

dependencies

Return the stacks this stack depends on

Declaration

readonly dependencies: Stack[];

Property Value

Stack[]

environment

The environment coordinates in which this stack is deployed. In the form aws://account/region. Use stack.account and stack.region to obtain the specific values, no need to parse.

You can use this value to determine if two stacks are targeting the same environment.

If either stack.account or stack.region are not concrete values (e.g. Aws.account or Aws.region) the special strings unknown-account and/or unknown-region will be used respectively to indicate this stack is region/account-agnostic.

Declaration

readonly environment: string;

Property Value

string

nested

Indicates if this is a nested stack, in which case parentStack will include a reference to it's parent.

Declaration

readonly nested: boolean;

Property Value

boolean

notificationArns

Returns the list of notification Amazon Resource Names (ARNs) for the current stack.

Declaration

readonly notificationArns: string[];

Property Value

string[]

parentStack

Returns the parent stack if this stack is nested.

Declaration

readonly parentStack?: Stack;

Property Value

Stack

partition

The partition in which this stack is defined

Declaration

readonly partition: string;

Property Value

string

region

The AWS region into which this stack will be deployed (e.g. us-west-2).

This value is resolved according to the following rules:

1. The value provided to env.region when the stack is defined. This can either be a concerete region (e.g. us-west-2) or the Aws.region token. 3. Aws.region, which is represents the CloudFormation intrinsic reference { "Ref": "AWS::Region" } encoded as a string token.

Preferably, you should use the return value as an opaque string and not attempt to parse it to implement your logic. If you do, you must first check that it is a concerete value an not an unresolved token. If this value is an unresolved token (Token.isUnresolved(stack.region) returns true), this implies that the user wishes that this stack will synthesize into a **region-agnostic template**. In this case, your code should either fail (throw an error, emit a synth error using node.addError) or implement some other region-agnostic behavior.

Declaration

readonly region: string;

Property Value

string

stackId

The ID of the stack

Declaration

readonly stackId: string;

Property Value

string

stackName

The concrete CloudFormation physical stack name.

This is either the name defined explicitly in the stackName prop or allocated based on the stack's location in the construct tree. Stacks that are directly defined under the app use their construct id as their stack name. Stacks that are defined deeper within the tree will use a hashed naming scheme based on the construct path to ensure uniqueness.

If you wish to obtain the deploy-time AWS::StackName intrinsic, you can use Aws.stackName directly.

Declaration

readonly stackName: string;

Property Value

string

tags

Tags to be applied to the stack.

Declaration

readonly tags: TagManager;

Property Value

TagManager

templateFile

The name of the CloudFormation template file emitted to the output directory during synthesis.

Declaration

readonly templateFile: string;

Property Value

string

templateOptions

Options for CloudFormation template (like version, transform, description).

Declaration

readonly templateOptions: ITemplateOptions;

Property Value

ITemplateOptions

templateUrl

An attribute (late-bound) that represents the URL of the template file in the deployment bucket.

Declaration

readonly templateUrl: string;

Property Value

string

urlSuffix

The Amazon domain suffix for the region in which this stack is defined

Declaration

readonly urlSuffix: string;

Property Value

string

Methods

addDependency(stack, reason)

Add a dependency between this stack and another stack

Declaration

addDependency(stack: Stack, reason?: string): void;

Parameters

stack Stack

reason string

Returns

void

addDockerImageAsset(asset)

Declaration

addDockerImageAsset(asset: DockerImageAssetSource): DockerImageAssetLocation;

Parameters

asset DockerImageAssetSource

Returns

DockerImageAssetLocation

addFileAsset(asset)

Declaration

addFileAsset(asset: FileAssetSource): FileAssetLocation;

Parameters

asset FileAssetSource

Returns

FileAssetLocation

allocateLogicalId(cfnElement)

Returns the naming scheme used to allocate logical IDs. By default, uses the HashedAddressingScheme but this method can be overridden to customize this behavior.

In order to make sure logical IDs are unique and stable, we hash the resource construct tree path (i.e. toplevel/secondlevel/.../myresource) and add it as a suffix to the path components joined without a separator (CloudFormation IDs only allow alphanumeric characters).

The result will be:

<path.join('')><md5(path.join('/')> "human" "hash"

If the "human" part of the ID exceeds 240 characters, we simply trim it so the total ID doesn't exceed CloudFormation's 255 character limit.

We only take 8 characters from the md5 hash (0.000005 chance of collision).

Special cases:

• If the path only contains a single component (i.e. it's a top-level resource), we won't add the hash to it. The hash is not needed for disamiguation and also, it allows for a more straightforward migration an existing CloudFormation template to a CDK stack without logical ID changes (or renames). - For aesthetic reasons, if the last components of the path are the same (i.e. L1/L2/Pipeline/Pipeline), they will be de-duplicated to make the resulting human portion of the ID more pleasing: L1L2Pipeline<HASH> instead of L1L2PipelinePipeline<HASH> - If a component is named "Default" it will be omitted from the path. This allows refactoring higher level abstractions around constructs without affecting the IDs of already deployed resources. - If a component is named "Resource" it will be omitted from the user-visible path, but included in the hash. This reduces visual noise in the human readable part of the identifier.

Declaration

protected allocateLogicalId(cfnElement: CfnElement): string;

Parameters

cfnElement CfnElement

Returns

string

formatArn(components)

Creates an ARN from components.

If partition, region or account are not specified, the stack's partition, region and account will be used.

If any component is the empty string, an empty string will be inserted into the generated ARN at the location that component corresponds to.

The ARN will be formatted as follows:

arn:{partition}:{service}:{region}:{account}:{resource}{sep}}{resource-name}

The required ARN pieces that are omitted will be taken from the stack that the 'scope' is attached to. If all ARN pieces are supplied, the supplied scope can be 'undefined'.

Declaration

formatArn(components: ArnComponents): string;

Parameters

components ArnComponents

Returns

string

getLogicalId(element)

Allocates a stack-unique CloudFormation-compatible logical identity for a specific resource.

This method is called when a CfnElement is created and used to render the initial logical identity of resources. Logical ID renames are applied at this stage.

This method uses the protected method allocateLogicalId to render the logical ID for an element. To modify the naming scheme, extend the Stack class and override this method.

Declaration

getLogicalId(element: CfnElement): string;

Parameters

element CfnElement

Returns

string

isStack(x)

Return whether the given object is a Stack.

We do attribute detection since we can't reliably use 'instanceof'.

Declaration

static isStack(x: any): x is Stack;

Parameters

x any

Returns

x is Stack

of(construct)

Looks up the first stack scope in which construct is defined. Fails if there is no stack up the tree.

Declaration

static of(construct: IConstruct): Stack;

Parameters

construct IConstruct

Returns

Stack

parseArn(arn, sepIfToken, hasName)

Given an ARN, parses it and returns components.

If the ARN is a concrete string, it will be parsed and validated. The separator (sep) will be set to '/' if the 6th component includes a '/', in which case, resource will be set to the value before the '/' and resourceName will be the rest. In case there is no '/', resource will be set to the 6th components and resourceName will be set to the rest of the string.

If the ARN includes tokens (or is a token), the ARN cannot be validated, since we don't have the actual value yet at the time of this function call. You will have to know the separator and the type of ARN. The resulting ArnComponents object will contain tokens for the subexpressions of the ARN, not string literals. In this case this function cannot properly parse the complete final resourceName (path) out of ARNs that use '/' to both separate the 'resource' from the 'resourceName' AND to subdivide the resourceName further. For example, in S3 ARNs:

arn:aws:s3:::my_corporate_bucket/path/to/exampleobject.png

After parsing the resourceName will not contain 'path/to/exampleobject.png' but simply 'path'. This is a limitation because there is no slicing functionality in CloudFormation templates.

Declaration

parseArn(arn: string, sepIfToken?: string, hasName?: boolean): ArnComponents;

Parameters

arn string

sepIfToken string

hasName boolean

Returns

ArnComponents

an ArnComponents object which allows access to the various components of the ARN.

prepare()

Prepare stack

Find all CloudFormation references and tell them we're consuming them.

Find all dependencies as well and add the appropriate DependsOn fields.

Declaration

protected prepare(): void;

Returns

void

prepareCrossReference(sourceStack, reference)

Exports a resolvable value for use in another stack.

Declaration

protected prepareCrossReference(sourceStack: Stack, reference: Reference): IResolvable;

Parameters

sourceStack Stack

reference Reference

Returns

IResolvable

a token that can be used to reference the value from the producing stack.

renameLogicalId(oldId, newId)

Rename a generated logical identities

To modify the naming scheme strategy, extend the Stack class and override the createNamingScheme method.

Declaration

renameLogicalId(oldId: string, newId: string): void;

Parameters

oldId string

newId string

Returns

void

reportMissingContext(report)

Indicate that a context key was expected

Contains instructions which will be emitted into the cloud assembly on how the key should be supplied.

Declaration

reportMissingContext(report: cxapi.MissingContext): void;

Parameters

report cxapi.MissingContext

Returns

void

resolve(obj)

Resolve a tokenized value in the context of the current stack.

Declaration

resolve(obj: any): any;

Parameters

obj any

Returns

any

synthesize(session)

Declaration

protected synthesize(session: ISynthesisSession): void;

Parameters

session ISynthesisSession

Returns

void

toJsonString(obj, space)

Convert an object, potentially containing tokens, to a JSON string

Declaration

toJsonString(obj: any, space?: number): string;

Parameters

obj any

space number

Returns

string