[question|feauture] metrics from /var/run/gluster/metrics

[pznamensky]

Glusterfs 4.0 brings tons of useful metrics, which can be read from /var/run/gluster/metrics after sending USR2 to gluster processes (https://docs.gluster.org/en/latest/release-notes/4.1.0/#monitoring). Is it planning to add those metrics into exporter? Would be very helpful for us.

[pranithk]

@pznamensky Over time yes. Not in the first release though

[amarts]

Is it planning to add those metrics into exporter?

I wrote something for grafana/graphite https://github.com/amarts/glustermetrics but for prometheus we need little more time to get there.

[pznamensky]

I ended up writing https://github.com/pznamensky/gmetrics-exporter, which is and nearly the same as https://github.com/amarts/glustermetrics but for Prometheus. Thanks, @amarts. And I still hope gluster/gluster-prometheus will do it inside own exporter.

[amarts]

Glad to see some effort on this.

@pznamensky how is it working for you? Can you write a small blog (with may be 1 or 2 screenshare) on it? That would be amazing.

Also note that, monitoring is something many users are looking at, so I believe talking about it in different places, would help getting users to your repo :-) (or we may consider importing your repo into gluster/ org in github some day)

[pznamensky]

After installing the exporter, we got an overview of the load type per client. It is really helpful for us. And that's why I hope Gluster-prometheus project will include similar feature to its exporter. But as for me, the best option is to include exporter directly to FUSE client. I probably write a small blog later.