PMM 2.32 - Identify unhealthy targets

Hi,

We forget to restart the agents on on our database servers sometimes after the outage. is there a way to alert us if the agents are not connected to the configured systems in PMM.

from PMM inventory , we have to browse through to check all the agents that are with status UNKNOWN.

Prevously we were able to check using the below link. that links displayed below are no longer working.

http://pmm-host/prometheus

Single-node VictoriaMetrics

See docs at https://docs.victoriametrics.com/
Useful endpoints:
vmui - Web UI
targets - discovered targets list
api/v1/targets - advanced information about discovered targets in JSON format
config - -promscrape.config contents
metrics - available service metrics
flags - command-line flags
api/v1/status/tsdb - tsdb status page
api/v1/status/top_queries - top queries
api/v1/status/active_queries - active queries
/graph/explore - explore metrics grafana page
/graph/d/prometheus-advanced/advanced-data-exploration - PMM grafana dashboard

1 Like

I believe you are looking for https://pmm-host/victoriametrics/targets

and you can check “unhealthy” to filter by those hosts. I bet you could also create an alert to also let you know via email/slack or just the firing alerts.

1 Like

Not really. It is only show extractors with errors.

We have system connected with pmm agents. Let’s suppose the pmm agent is down.

We see lot of agents with unknown status in PMM inventory page.

Currently I don’t know of a way to check them instead of doing a browser search for unknown on every page of agents in pmm inventory

We also like to easily spot them and also alert on them

1 Like