Step 3: Create the console bucket

To use the Amazon Lookout for Vision console, you need an Amazon S3 bucket that is known as the console bucket. The console bucket stores the following:

• Images that you upload to a dataset with the console.

• Training results for model training that you start with console.

• Trial detection results.

• Temporary manifest files that the console creates when you use the console to create a dataset by automatically labeling images in an S3 bucket. The console doesn't delete the manifest files.

When you first open the Amazon Lookout for Vision console in a new AWS Region, Lookout for Vision creates the console bucket on your behalf. Note the console bucket name because you might need to use the bucket name in AWS SDK operations or console tasks, such as creating a dataset.

Alternatively, you can create the console bucket by using Amazon S3. Use this approach if Amazon S3 bucket policies don't let the Amazon Lookout for Vision console successfully create the console bucket. For example, a policy that disallows the automatic creation of an Amazon S3 bucket.

Note

If you only use the AWS SDK and not the Lookout for Vision console, you don't need to create the console bucket. You can use a different S3 bucket with a name of your choosing.

The format of the console bucket name is lookoutvision-<region>-<random value>. The random value ensures that there isn't a collision between bucket names.

Creating the console bucket with the Amazon Lookout for Vision console

Use the following procedure to create the console bucket for an AWS Region with the Amazon Lookout for Vision console. For information about the S3 bucket settings that we enable, see Console bucket settings.

To create the console bucket by using the Amazon Lookout for Vision console

1. Ensure that the user or group you are using has AmazonLookoutVisionConsoleFullAccess permission. For more information, see Step 2: Set up permissions.

2. Open the Amazon Lookout for Vision console at https://console.aws.amazon.com/lookoutvision/.

3. On the navigation bar, choose Select a region. Then choose the AWS Region for which you want to create the console bucket.

4. Choose Get started.

5. If this is the first time that you've opened the console in the current AWS Region, do the following in the First time set up dialog box:

   1. Copy down the name of the Amazon S3 bucket that's shown. You'll need this information later.

   2. Choose Create S3 bucket to let Amazon Lookout for Vision create the console bucket on your behalf.

   The First time set up dialog box is not shown if the console bucket for the current AWS Region already exists.

6. Close the browser window.

Creating the console bucket with Amazon S3

You can use Amazon S3 to create the console bucket. You must create the bucket with Amazon S3 versioning enabled. We recommend that you use an Amazon S3 lifecycle configuration to remove noncurrent (previous) versions of an object and delete incomplete multipart uploads. We don't recommend a lifecycle configuration that deletes current versions of an object. For information about the S3 bucket settings we enable for console buckets that you create with the Amazon Lookout for Vision console, see Console bucket settings.

1. Decide the AWS Region in which you want to create a console bucket. For information about supported regions, see Amazon Lookout for Vision endpoints and quotas.

2. Create a bucket using the S3 console instructions at Creating a bucket. Do the following:

   1. For step 3 specify a bucket name that is prepended with lookoutvision-region-your-identifier. Change region to the region code you chose in the previous step. Change your-identifier to a unique identifier of your choosing. For example, lookoutvision-us-east-1-my-console-bucket-1

   2. For step 4 choose the AWS Region that you want to use.

3. Enable versioning for the bucket by following the S3 console instructions at Enabling versioning on buckets.

4. (Optional) Specify a lifecycle configuration for the bucket by following the S3 console instructions at Setting lifecycle configuration on a bucket. Do the following to remove noncurrent (previous) versions of an object and delete incomplete multipart uploads. You don't need to do steps 6, 8, 9, 10.

   1. For step 5 choose Apply to all objects in the bucket.

   2. For step 7 select Permanently delete noncurrent versions of objects and Delete expired object delete markers or incomplete multipart uploads.

   3. For step 11 enter the numbers of days to wait before deleting noncurrent versions of an object.

   4. For step 12 enter the number of days to wait before deleting incomplete multipart uploads.

Console bucket settings

If you create the console bucket with the Amazon Lookout for Vision console, we enable the following settings on the console bucket.

• Versioning of objects in the console bucket.

• Server-side encryption of objects in the console bucket.

• A lifecycle configuration for the deletion of noncurrent objects (30 days) and incomplete multipart uploads (3 days).

• Block public access to the console bucket.