The TimeMonitorManager only reports the first error when querying the time synchronization. This is by design as the application runs often we don't want to flood the log. Basically, only configuration changes and state changes should be logged. However, this makes harder to troubleshoot issues with the NTP server address for instance. As only the first error is reported changing to a different (but still incorrect) address won't show anything in the log. It would be better if we did log the error if it is a different message. Note, that we need to be careful to only consider parts of the error message that does not change (excluding the npdate PID or any time-stamps).
- Add unit test
Way to reproduce (from head so might not be 100% accurate):
1. Configure timeserver.hosts worker property with an IP address that is not an NTP server
2. See the error message in the log about that IP address
3. Change timeserver.hosts worker property to a different IP address that is also not an NTP server
Expected: One new error message about the different IP address
Actually: No new error message