Serverless Stack

Hello, I’m using `Script` inside my db stack to run a migration when the construct is deployed.
When I deploy or remove the stack using `sst start` / `sst remove`  it takes too much time. For example when I run a remove, CloudFormation takes about 40 minutes to switch from the status `MigrationScriptonCreateFunction DELETE_IN_PROGRESS` to the `MigrationScriptonCreateFunction DELETE_COMPLETE` . This seems to happen only with Lambda functions managed by `Script` , other Lambdas delete smoothly.

this usually is an indicator the script is erroring when being brought down

it's really annoying but cfn does not deal with custom resource errors gracefully, it waits for some sort of large timeout

<@U03CDLDPF41> if you go to CloudWatch and find the log group for the Script’s Lambda function. Do you see errors in there?

I also thought the problem could be the script erroring, but I can’t figure out why the script is called also during the removal… :slightly_smiling_face: (I’m talking about `onCreate`  script)

<@U01JVDKASAC> no error within CloudWatch, neither within CloudTrails events…