Currently the supervisor logs runnable restarts into its own supervisor DN. This hides runnable restarts from DN-filtered logs, potentially confusing people who look at the logs as a lot of our runnables do not log errors, instead relying on the supervisor to log them as the reason for restarting the runnable. It would be a much better experience if these were inline with the actual runnable logs. For example #295 can only really be understood by looking at both the original DN logs as well as the supervisor's.
Currently the supervisor logs runnable restarts into its own
supervisor
DN. This hides runnable restarts from DN-filtered logs, potentially confusing people who look at the logs as a lot of our runnables do not log errors, instead relying on the supervisor to log them as the reason for restarting the runnable. It would be a much better experience if these were inline with the actual runnable logs. For example #295 can only really be understood by looking at both the original DN logs as well as the supervisor's.