Hi,
we want to create a plugin to monitor our ceph radosgw period and would like to use values that come from the agent (via agent plugin) in other hosts.
The idea behind that:
- collect the period map from the host which contains several cluster values like period UUID, hostnames, endpoints and epochs.
- create a check for the period ID that validates if the host got the correct epoch.
- the epoch is a simple integer which increments if a cluster does a change to the period and commits it (commiting means it will get synced to the master and all other clusters fetch that.
Is there a way to check for the largest epoch in a period and use this as the “ok” value for all checks that uses this period?
Every cluster contains three hosts and a period can have 1…N clusters, so this might add up in hosts.
Alternativly I would like to have a dummy host that will use the period map of all hosts and build a service map that reflects the cluster-period combination. This host would also validate SSL (lifetime and chain) and if a http(s) connect is possible (result should not be a 5xx)
Is this possible?