Why does "Top Indexes for Database (XX)" periodically go into unknown state?

Question

I believe this might be similar to APPINSIGHT for SQL - Unknown status for Top Indexes .etc but it has nothing to do with the logon setting.  Essentially, on a very regular basis the Top Indexes component monitor goes into an unknown state for a select few databases, and it's the only component monitor that behaves this way (at least with such regularity).  I'm polling every 5 minutes, so getting the error exactly at 1 hour intervals for any given database is a bit baffling.  I don't suppose anyone can shed light on exactly what query this monitor is executing and/or why this clockwork outage might be happening?

Petr.Vilem · Answer

Just make sure to increase timeout for whole application accordingly when you increase timeout(s) for individual component(s). Indexes are being polled with every 12th poll (with default values of 1 hour and 5 minutes) but it is still happening within main application polling job.

tomasv · Answer

Since SAM 6.2.4 there is setting for timeout of the Top indexes query.

If you have issues with it, you can increase the timeout.

This statistics has it's own polling frequency - default  value of it is1 hour (3600 sec).

So the timeout can be set to value like 20 minutes (1200 sec).

aLTeReGo · Answer

You are seeing this occur once an hour because index fragmentation is not collected at the same 5 minute interval that performance/availability data is.

The unknown state is likely the result of this polling failing to complete successfully, likely because it is running out of time. If that is the issue, you can resolve it by doubling the default timeout value for this application.