You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Patroni has a port 8008 on localhost in the container, on this port you can find metrics on the /metrics endpoint and status of the working of the cluster on /cluster (just tested this inside the pod with curl -k https://localhost:8008/metrics). I don't see a way to enable the exposure of this port on the container.
Use Case
The endpoint https://127.0.0.1:8008/cluster gives an overview of the current status of the cluster, including the replication lag. We can use this endpoint to monitor if the cluster is healthy, because if the lag is to high and the master node goes down. patroni can't switchover to the replica with the high lag. A possible scenario where the whole cluster goes down. Based on that endpoint we can also create alerting rules if the lag is to high.
I can create a feature request to have this port exposed in our backlog. Can you tell me a bit more about your use case?
We actually have alerting and trending of replication lag in our monitoring stack, so you are already covered there. Are there other metrics that these endpoints provide that you are particularly interested in?
I want to monitor the total current lag of patroni, As far as i can see in the replication lag in the monitoring stack doesn't show the total lag, i only saw spikes on times where there was a bit of lag in that dashboard.
Overview
Patroni has a port 8008 on localhost in the container, on this port you can find metrics on the
/metrics
endpoint and status of the working of the cluster on/cluster
(just tested this inside the pod withcurl -k https://localhost:8008/metrics
). I don't see a way to enable the exposure of this port on the container.Use Case
The endpoint
https://127.0.0.1:8008/cluster
gives an overview of the current status of the cluster, including the replication lag. We can use this endpoint to monitor if the cluster is healthy, because if the lag is to high and the master node goes down. patroni can't switchover to the replica with the high lag. A possible scenario where the whole cluster goes down. Based on that endpoint we can also create alerting rules if the lag is to high.Besides that we can use the metrics endpoint for metrics and we can use an Grafana dashboards to give a clear insight on the status of Patroni. With example this dashboard: https://grafana.com/grafana/dashboards/18870-postgresql-patroni/.
Desired Behavior
A possiblity to enable the exposure of the port on the postgrescluster
Think about (or someting else):
postgrescluster.spec.patroni.enableport
: trueor enable this port default in every Postgrescluster.
The text was updated successfully, but these errors were encountered: