A scheduled job that normally runs in under 10 seconds will suddenly take hours to complete. The amount of work the job can do is fixed and the issue is not load related. Checking the status of the SPID executing the job it appears to stay in a runnable status most of the time. CPU utilization on the server is < 35%, 24 core with 265GB memory. This jobs runs every minute and except for the first few runs after it hangs which then can take 2-3 minutes to complete.
his job executes a stored proc, and pretty much fully tuned. This stored proc runs without issues in a lot of other sites with less hadrware with out issue. So where do I go looking next?