Metrics and monitoring

[jlebon]

We will soon have two services here: coreos-koji-tagger and config-bot. We need to figure out how we'll monitor these services and e.g. be alerted if they go down.

Right now coreos-koji-tagger is planned to run in the OpenShift infra. Would probably make things easier to implement if we just run config-bot there too.

See related discussions in #3 (comment).

/cc @lucab since he has experience in that area (and I think was looking at this for the Cincinnati server as well?)

[lucab]

I think that coreos-koji-tagger, like the Cincinnati server, runs on Fedora infrastructure: https://infrastructure.fedoraproject.org/cgit/ansible.git/tree/playbooks/openshift-apps/coreos-koji-tagger.yml

To my knowledge, that Openshift cluster does not (yet) have Prometheus+AlertManager, but monitoring is done via Nagios.