Retry TestCustomAuthorizerApp deployment on transient IAM role propagation failure

[GarrettBeatty]

Problem

The durabletesting3 CI job (run 28258443522) failed in the TestCustomAuthorizerApp.IntegrationTests project — all 20 tests failed because the fixture's CloudFormation deployment rolled back:

SimpleHttpApiUserInfo   CREATE_FAILED
Resource handler returned message: "The role defined for the function cannot be assumed by Lambda.
(Service: Lambda, Status Code: 400 ...)" (HandlerErrorCode: InvalidRequest)

This is a transient IAM eventual-consistency race: the stack creates a per-function IAM role and immediately calls Lambda CreateFunction, but the role's trust policy hasn't propagated through IAM yet. Once one function fails, CloudFormation cancels the other in-flight resources and rolls the whole stack back (ROLLBACK_COMPLETE), failing every test in the project. It is unrelated to the PR's code changes and typically passes on re-run.

Fix

Wrap dotnet lambda deploy-serverless in DeploymentScript.ps1 with a retry loop (3 attempts):

• Between attempts, delete the rolled-back stack — a ROLLBACK_COMPLETE stack cannot be updated or re-created — and wait stack-delete-complete before retrying.
• Add a brief pause to give IAM additional time to settle.
• Surface CloudFormation failed-resource events on each failure for easier debugging.

This makes the integration test resilient to the transient IAM propagation failure instead of failing the whole CI job.

Testing

• PowerShell script parses cleanly ([Parser]::ParseFile).
• The retry path only triggers on deployment failure; the happy path is unchanged (single deploy, then break).