Is there a sample showing how to implement a fault-tolerant workflow that lives longer than the retention time of the logs?
I’m trying to understand whether this is a job for something like continueAsNew ahead of the log expiration time or whether the workflow needs to use an Activity to push its state somewhere to manually ensure it can restart after the original log data is gone.
Retention applies only to closed workflows. A workflow can stay open as long as needed. ContinueAsNew is needed only to limit history growth. If a workflow is long running but its history size is bounded, it can run for a very long time.