While trying to debug a singleton deployment test cases that very rarely intermittently fail, we realized we can't simply determine which node was the logging coming from. This is because the test log collates logs from all nodes into one log and since e.g. our main branch nightly builds do not preserver entire server logs, we were unable to debug the particular situation.

It comes to my mind that perhaps for clustering tests we can augment the CLI scripts that reconfigure all nodes to actually include the node name in the log.

This would only be needed for the console logging.