Ceph » Orchestrator

Once the host facts are part of the mgr/cephadm cache, he idea is to periodically check the host cache to look for anomalies. Any anomaly woul dbe flagged as a mgr healthcheck WARN alert.

These are the kinds of checks that could be implemented;

• OS version consistency
• Support consistently configured
• Security policy (SELINUX, AppArmor enabled) is consistent
• cluster network and public network using consistent * network connection speeds * interface MTU size * dedicated/shared interfaces

The longer term goal would be to allow these checks to be enabled/disabled from the UI.