We have some use cases where we leverage ApplicationFailure.nonRetryableError for some failures that we expect to happen from an activity.
In these cases, we don’t want these failures to be logged as error. Since we also use a service like Sentry for exception monitoring in our prod services, these expected activity failures will pollute the log and would prefer to only log unexpected failures.
These exceptions are logged from: io.temporal.internal.sync.POJOWorkflowImplementationFactory
It would be nice if we can customize the logging behavior for these failures. Perhaps a way to exclude certain error types (e.g. type field in ApplicationFailure) from being logged, or a callback that we can register to implement our own logging logic.
Additional context in Slack: Slack