Redfish sensors based alerts has no duration so could flapping when the value is fluctuating around max value.

[err404r]

I faced a two issues today first is the behavior similar to Issue#112 This time Fan readings are flapping from None to value and back, on the same server.

And this generates a huge amount of alerts. I will add a small duration to alert rule form 1 to 3 minutes probably.
To avoid false positives caused by flapping redfish api or something like that.

[jneo8]

close & re-open to trigger the github to JIRA action.

[err404r]

The same applies to RedfishStorageControllerHealthNotOk alerts and probably other alerts, we need take into account that readings from redfish are not perfectly stable

[err404r]

And to IPMISensorStateNotOk alert

[dashmage]

The Redfish "HealthNotAvailable" alerts have been removed since in a lot of cases the data might not be available but we wouldn't want an alert for each of them. The original linked issue has also been resolved.

@err404r - would this solve your issues as well?

[err404r]

Hi, no this issue is more general. then HealthNotAvailable.
On some motherboards readings are not 100% stable. So it's ok to periodically have sensor reading error.
I would say all sensor based alert rules should have for: 5m in their definition.
This will remove alerts from a single bad readings, short temperature spikes etc and make our life much more comfortable.

[dashmage]

Thanks for your comment and I understand your intention better. Let me check in with the rest of the team for their opinion and then pick a suitable duration time.