Hi @undone quiver is there a way or a plan to monitore the pollers if they are running? For the NAbox 2 we had like a check script running in the background which Prometheus could read.
#NAbox 4 - Monitore pollers
obtuse sedge
uneven sage
Does the Metadata Dashboard in Harvest help track this?
obtuse sedge
Definitly the best in our environemnt would be that these metric could be exposed like https://<nabox>vm/metrics so we could scrap the status and set alerts on it
but I'm not sure is this is possible in a easy way or Yann needs to implement something
uneven sage
You can try endpoint like https://<nabox>/vm/api/v1/query?query=poller_status
https://netapp.github.io/harvest/24.11.1/ontap-metrics/#poller_status
But usually some sort of alertmanager type of configuration is better for such tasks. For vm it is https://docs.victoriametrics.com/operator/resources/vmalertmanager/
Please note that the poller_status metric is only available for pollers that are currently running. If a poller stops or is not in the start state, this metric will not be available.
obtuse sedge
Absolutely we have an external Prometheus and Alertmanager in place for the whole alarming. I just know in the past with Harvest 1.x and NAbox 2 sometimes the pollers crashes and we where not aware of that. To be honest until now I did not saw such issues again with the new versions of both 😊
I‘ve just counted we have currently like 12 NAbox instances with around 90 pollers in place and I would be nice if there is a way to monitore all the pollers, just in case one of them would die (for whatever reason)
uneven sage
We probably would need a centralised prometheus or vm from where all of this can be monitored.
obtuse sedge
A central instance we already have and that collects the metrics from the onboard node exporter. Probably Yann can extend this with the poller status 🙈