We would like a custom alert that looks for single queries that generate more than a given amount of wait time within a certain time period (for example, 20 minutes of total wait time within 10 minutes). This would allow catching either single long-running queries as well as shorter queries that are being executed with great frequency and potentially generating undue load on the server.
I see slightly different examples for total wait time for the whole server and for long-running queries, but not for this variation. The target platform is SQL Server.
Any thoughts?