Details
-
Bug
-
Resolution: Done
-
Major - P3
-
None
-
None
-
ALL
Description
When we set a node as "down" due to a failed heartbeat, we currently log nothing at debug log level 0. It's hard to tell why a primary stepped down, for instance.
We should log when we take note of a change in a remote node's state. I would also consider logging when heartbeats fail, and the reason why (TCP disconnect vs. timeout).