cclark-1005023 (3/24/2010)
I like the thoughts in this, but we use a more general solution where we track similar information for all production jobs. Instead of using a manually configured time period for each job we calculate the standard deviation based on the historical job durations and report only when the current run time falls outside of 3 standard deviations. This alerts us to both jobs that run too long or too short, etc.
Nice Idea. Since, statistically, 99% should fall within that range of deviations - it should be more accurate to find when a job is behaving abnormally and avoid false positives.
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events