What is the problem?

A deployment failed for our team using a CDK defined infrastructure package due to a failure in the bucket notification lambda function. After looking at the logs in our account, it appears that it was caused due to missing error handling within the bucket notification lambda function. The line that failed was this one, which is called here.

Error logs:

[ERROR] 2021-10-05T18:18:42.172Z 5948f640-82f4-44de-a218-694408fb7bfb Failed to put bucket notification configuration

put_bucket_notification_configuration(bucket, config)

botocore.exceptions.ClientError: An error occurred (OperationAborted) when calling the PutBucketNotificationConfiguration operation: A conflicting conditional operation is currently in progress against this resource. Please try again.

Reproduction Steps

Have multiple notifications attempt to update the bucket? I am not quite sure!

What did you expect to happen?

The bucket notification succeeds. To do this, adding an error handling + retry method here would be great. For transient errors like this, it seems like having more robust error handling and retrying transient errors would be very beneficial. That way all of the consumers of AWS CDK get a smoother experience here

What actually happened?

The bucket notification failed, causing the stack to get into an UPDATE_ROLLBACK_FAILED state

CDK CLI Version

1.25.0

Framework Version

n/a

Node.js Version

n/a

OS

linux

Language

Python

Language Version

3.8

Other information

List of events: