We have an availability group
Replica-01 (Sync on-premise) PRIMARY Replica-02 (Sync on-premise) SECONDARY Replica-03 (Async Azure VM) SECONDARY
Replica-01 is sometimes having lease time out and availability group is entering RESOLVING state for short period of time, effectively an outage on production for 30-60 sec
Looking at Cluster log, we see below:
Lease renewal failed because the existing lease is no longer valid. The lease is expired
Availability Group lease is no longer valid. Resource Alive result 0.
Cluster log reports that performance counters collected before lease time out showed 100% CPU utilization
Some server-wide event is happening that is causing Availability Group / Primary replica loosing heartbeat with Windows Failover Cluster, but what exactly ?
How can I diagnose and resolve those type of problem ?