Update ResourceNotificationServiceLogging#4693
Conversation
- Only log state text changes at debug level when the state text actually changes - Log all published updates at the trace level
|
Is there any chance that we'll get two resource notifications for a resource being processed in parallel. I'm wondering if there is potential for resource states to rapidly oscillate and then we don't fire them. |
The publish method takes a lock on the state, so all updates for a resource get effectively sequentialized anyway, so while races might result in "extra" transitions being logged, I'm not sure any can be missed, e.g.
If in step 4 state2 update gets the lock first, then state1 to state2 transition is logged, and then state1 update logs transition from state2 to state1 (the "extra" update). If we wanted to solve that we could use a |
Fixes #4686
Microsoft Reviewers: Open in CodeFlow