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
In my case, that happens to be a smaller machine than other peers, and running the exporter with all collectors enabled generates a visible load (and audible fan speed increase).
(TBH, my home cluster is using small hosts and that's most probably not a recommended sizing).
It would be great to be able to control on which host cluster metrics are exported.
For example, using a configuration flag : expose-cluster-metrics: [auto|true|false] auto would be the default and would still use the current IsLeader test. true and false would override the IsLeader test.
The text was updated successfully, but these errors were encountered:
Currently, cluster metrics are only exported on the Leader node.
For a GD1 cluster, that's the peer with the maximum UUID (lexicographically) (https://github.com/gluster/gluster-prometheus/blob/master/pkg/glusterutils/exporterd.go#L23)
In my case, that happens to be a smaller machine than other peers, and running the exporter with all collectors enabled generates a visible load (and audible fan speed increase).
(TBH, my home cluster is using small hosts and that's most probably not a recommended sizing).
It would be great to be able to control on which host cluster metrics are exported.
For example, using a configuration flag :
expose-cluster-metrics: [auto|true|false]
auto
would be the default and would still use the current IsLeader test.true
andfalse
would override the IsLeader test.The text was updated successfully, but these errors were encountered: