I’m having this problem with Eaton 5P 1500RT UPS where the Health Inverter Status keeps getting turned back on, seems to happen after a week or so. Not sure how to stop this as it’s not an alarm I’m concerned with?
I saw from this post a very similar issue, but the resolution didn’t seem clear to me Disabled sensor alerts keep getting re-enabled automatically
Why would the sensors be deleted and re-created?
I had a similar issue but I wasn’t sure if it was my own ineptitude with LibreNMS or something wasn’t right.
I get an alarm for voltage on a transceiver. Initially I tried altering the thresholds for the port in question, I still received alerts. I tried going into the device itself and shutoff monitoring of the sensor in question, I kept getting alerts. Eventually I went to the alert itself and added the device in question to the exemptions. Not really graceful at all but it works. Only downside is that if another sensor alarm is tripped on that switch I’m not going to get an alert for it.
In our case, the SNMP rewalks of the device were failling sometimes due to poor network infrastructure (dropped UDP reply packets etc.) I’m guessing at this point, any configuration info associated with the disappeared interfaces (e.g. whether to alert or not) would be removed from Libre too.
This would then be followed at some point by a successful rewalk, which would reinstate the interfaces, but Libre was seeing them as new - it had no record of how I’d configured them previously.
Hope that helps …
Maybe, that makes some sense as this is a cell tower site where the wireless PTP occasionally has connection drop issues, but it only seems to affect the Eaton UPS and none of our routers in the hut monitored via SNMP.
In any case it seems like a bug…a rewalk shouldn’t set values back to default if UDP packets drop in an already discovered device.
Try increasing the snmp timeout for that device.
You can do it editing device -> SNMP
Thanks, I’ll give that a try. Didn’t realize I could do that at the per device level.