We've had several instances where our AppInsight for SQL instances go into an Unknown status; this occurs due to timeouts with the "Top Indexes for Database" on TempDB.
Looks like the query behind that is timing out periodically, even when the timeout value is increased.
The feature request is to either exempt TempDB from this monitor, or allow it to be disabled on a per-database option.
Given the highly transient nature of tables and indexes in TempDB, monitoring this seems pretty pointless, especially as it's throwing off the application status.
