Add metrics and alerts for basic connectivity between ceph nodes and cloudvirts, specifically:
- mons <-> osds (non jumbo frames, public interface only)
- osds <-> osds (jumbo frames, public and cluster interfaces)
- cloudvirts -> osds (non jumbo frames, public interface only) (will change to jumbos once T330075: [cloudvirt] Move to jumbo frames is there)
The alerts should not page, but just show up in alertmanager when:
- More than 4 pings from a node failed in a short period
- More than 4 pings to a node failed in a short period
- More than 10% of the pings in total failed in a short period
- More than 2 pings from a node failed continuously in a medium period
- More than 2 pings to a node failed continuously in a medium period
- More than 5% of the pings in total failed in a medium period