Cluster failovers should provide additional information

by Mar 1, 2017

When we have a cluster failover, sqldm can’t tell if it was user-initiated (failed over by an administrator as part of maintenance, for example) or due to a crash (unexpected crash/loss of cluster resource). By scanning the sql log, sqldm should be able to tell if this failover was expected and alert differently (or include notes in the alert as relevant)


PH22 over 6 years ago
For Always On Availability Groups too

Tep Chantra over 2 years ago
I know that the Failover alert specifically alerts when a failover happens but doesn’t provide any additional details. But SQLDM does have the SQL Server Error Log metric which scans the SQL Server error log for specific log entries (in some cases you have to edit the advanced options). But if you received a failover event, would it be OK if SQLDM also provided the SQL Server Error Log alert. Then use the two alerts together to determine whether it was something that you needed to look into further?

Just to give an example, let’s say a planned failover only generated the Failover Alert but an unexpected failover would generate both the Failover and SQL Server Error Log alert.