Grafana-Prometheus Dashboards using 'unpoller' in all metrics

[DrewBiermann]

Grafana-Prometheus Dashboards using 'unpoller' in all metrics but the metric data pulled in from the controller via prometheus is labeled 'unifipoller'.

E.g. unpoller_client_uptime_seconds vs unifipoller_client_uptime_seconds

Dashboards show up empty unless you go in and modify this.

[RedEchidnaUK]

Not sure if you fixed this, but I found the same issue. What confused me, I had it was working using a local install, then I ran it in a Docker container using the community Docker compose example and I noticed all of the metrics started with unifipoller.

I realised that the community Docker compose example sets the UP_PROMETHEUS_NAMESPACE=unifipoller, and that's the issue, it should be UP_PROMETHEUS_NAMESPACE=unpoller

[DrewBiermann]

@RedEchidnaUK I didn't catch that initially. Looking at my compose file now it does have it set to unifipoller. I wish I had seen that before I went and edited all the queries in the dashboards ha ha. Your solution is obviously the quickest fix.

[DrewBiermann]

@RedEchidnaUK has the easiest fix. Just modify your docker-compose to us UP_PROMETHEUS_NAMESPACE=unpoller and the dashboards will work.