Implement Prometheus metrics

[bsctl]

Currently, sending USR1 signal to keepalived process will dump configuration data to /tmp/keepalived.data, and sending USR2 will write statistics to /tmp/keepalived.stats inside the container.

$ kubectl -n kube-system exec kube-keepalived-master-0 -- kill -USR1 1
$ kubectl -n kube-system exec kube-keepalived-master-0 -- cat /tmp/keepalived.stats

It would be nice to have a sidecar container exposing statistics as Prometheus metrics.

[tuxpeople]

I recently stumbled upon this: https://github.com/cafebazaar/keepalived-exporter not sure if this would be feasible for this purpose.

[bsctl]

Hi @tuxpeople thanks for the suggestion but I prefer to stay away from GPL licensed code.