NewContinueAsNewError and retry policy propagation

If workflow is created and started with retry policy defined retry policy will be recorded in history and will be respected for the first run.

But if this workflow is using NewContinueAsNewError with the same context, in subsequent runs retry policy will NOT be propagated and in case of timeouts entire workflow will never be retried. Also retry policy doesn’t show in history of ContinueAsNew runs.

Does retry policy need to be explicitly set on context for each call to NewContinueAsNewError or is this a bug?

Seems it’s neither propagated nor can it be explicitly set. I have opened RetryPolicy not propagated on ContinueAsNew · Issue #695 · temporalio/sdk-go · GitHub. I will confirm which one we prefer (implicit or explicit) and solve in that issue.

Thank you Chad.

If this change goes through, can you estimate what release version would this be a part of, and when will it go out?

PR is open at Propagate retry policy for continue-as-new by cretz · Pull Request #696 · temporalio/sdk-go · GitHub. I think this would be part of next release which is early February.