Job duration taking too long - triggering WHILE job is still running
CONTEXT: If a job is taking a long time (outside of some predefined threshold), we want to be alerted about it. That’s often an actionable alert – we could decide to stop the job, we could let the users know that report XY will be late as a consequence etc. One step further is being alerted on the job STEP level as well.
PROBLEM: Current alert “Job duration unusual” is triggered AFTER the job completes, if it meets the alert's criteria. We need an alert to trigger WHILE the job (or job step) is still running, but the criteria is already met.
EXAMPLE SOLUTIONS:
• Change the “Job duration unusual”
• New alert(s)
Blaž Dakskobler
shared this idea
As of v10.0.4 the Job Duration Unusual alert is raised while jobs are still running, so marking this as completed. Thank you very much for the feedback.
-
As of v10.0.4 the Job Duration Unusual alert is raised while jobs are still running, so marking this as completed. Thank you very much for the feedback.
-
Blaž Dakskobler
commented
@Redgate: What's the status here?
-
Blaž Dakskobler
commented
@RedGate: consider merging these votes to https://sqlmonitor.uservoice.com/forums/91743-suggestions/suggestions/33919051-job-duration-taking-too-long-triggering-while-jo , which currently has 51 votes already.
-
Jay Graves
commented
I found a custom metric for a single long running job. However, what we need is a way to alert for SQL Agent jobs that are running beyond a designated time (similar to the long running query metric built in). The "unusual duration" alert doesn't quite do what we need. Hopefully someone knows how to make this a custom metric. If not, consider a feature request!
-
Thanks Blaž, that makes sense - I'll merge the suggestions and discuss this with the team
-
Blaž Dakskobler
commented
Please move these votes over to https://sqlmonitor.uservoice.com/forums/91743-suggestions/suggestions/33919051-job-duration-taking-too-long-triggering-while-jo , which currently has 44 votes already.
-
Robert
commented
Currently I don't have an option to get alerted if we want to know if a job is hung, been running for a long time and has been stuck?
SQL Monitor only offers the option to get alerted AFTER a job is finished with an unusual duration.
In case I have a job running usual a few minutes, but sometimes stuck during the execution which tends to much longer runtime. This won't be spotted by SQL Monitor.
Please, add a new alert to address such issues.
-
Anonymous
commented
Or you could get a better scheduler such as JAMS from MVPSI, which can send alerts on 'over run', or take some other action (like kill the job).
