Manage Executions¶

Use the AWS CLI to inspect, stop, update, and clean up durable functions and their executions.

List executions¶

aws lambda list-durable-executions-by-function \
  --function-name my-durable-function

Get execution details¶

aws lambda get-durable-execution \
  --durable-execution-arn <execution-arn>

Get execution history¶

View the checkpoint history for an execution:

aws lambda get-durable-execution-history \
  --durable-execution-arn <execution-arn>

Stop an execution¶

aws lambda stop-durable-execution \
  --durable-execution-arn <execution-arn>

Stopping an execution affects the durable execution and the running Lambda invocation differently. Understanding the distinction helps you predict what happens to code that is already running.

What stopping does¶

StopDurableExecution stops the durable execution. Once stopped, the execution no longer accepts checkpoints, nor does it start new Lambda invocations. Any operation that was in progress stays in the STARTED state, and no further terminal checkpoints (SUCCEEDED or FAILED) are recorded for it.

Stopping the execution does not stop the Lambda invocation that is currently running. Lambda has no mechanism to interrupt a function that is already executing, so any code that is mid-flight (for example, inside a step or a child context) keeps running.

What happens to running code¶

If your function is actively running when the execution is stopped, it continues until it reaches the next checkpoint. At that point the SDK detects that the execution has been stopped, raises an invocation error, and ends the current invocation. In practice this is usually at the end of the current block of code, when the SDK tries to checkpoint that block's result.

Update function code¶

After updating your code, publish a new version and point your alias to it.

aws lambda update-function-code \
  --function-name my-durable-function \
  --zip-file fileb://function.zip

aws lambda wait function-updated \
  --function-name my-durable-function

aws lambda publish-version \
  --function-name my-durable-function

aws lambda update-alias \
  --function-name my-durable-function \
  --name prod \
  --function-version <new-version>

aws lambda update-function-code \
  --function-name my-durable-function \
  --image-uri 123456789012.dkr.ecr.us-east-1.amazonaws.com/my-durable-function:latest

aws lambda wait function-updated \
  --function-name my-durable-function

aws lambda publish-version \
  --function-name my-durable-function

aws lambda update-alias \
  --function-name my-durable-function \
  --name prod \
  --function-version <new-version>

Running executions will continue to use the version they started with. New invocations use the updated alias.

If you're still actively developing and you don't want to publish a new version each time you update, you could use LATEST$ just during development, but please be very aware that executions might not replay correctly (or even fail) if the underlying code changes during running executions. Always use numbered versions or aliases in production.

View logs¶

aws logs tail /aws/lambda/my-durable-function --follow

Delete durable functions¶

aws lambda delete-function --function-name my-durable-function

aws iam detach-role-policy \
  --role-name durable-function-role \
  --policy-arn arn:aws:iam::aws:policy/service-role/AWSLambdaBasicDurableExecutionRolePolicy

aws iam delete-role --role-name durable-function-role

If you deployed as a container image, also delete the image from ECR:

aws ecr batch-delete-image \
  --repository-name my-durable-function \
  --image-ids imageTag=latest

Replace latest with the tag you pushed if you used a different tag. To delete multiple tags at once, specify each with a separate imageTag= argument:

aws ecr batch-delete-image \
  --repository-name my-durable-function \
  --image-ids imageTag=latest imageTag=v1.0.0