I am reporting on batch processing.
At the highest level there us the concept of a "Batchid" and within each batchid there can be multiple "Jobid"'s.
Each job can possible fail which provides a literal "error" in a field I interrogate.
I want to alert based on every time a unique BatchID has an "error" field associated.
However, I want to check every 10 minutes and only send an alert when the batchid and jobid's are different.
If I just supress based on the same batchid being present, I might miss the case where the origonal jobid gets resoved, but the next job under the same batchid fails.
This is what I want to do:
batchid jobid action
1234 432 alert
1234 432 throttle wait
1234 567 alert
1267 123 alert
Is this what you're referring to? I think you can put multiple fields in that throttle field list. So you can put it batchid and jobid in there maybe?
http://docs.splunk.com/Documentation/Splunk/latest/Alert/ThrottleAlerts
If that doesn't work or isn't an option, what about just eval'ing a new field in your search and then throttle based on it?
... | eval throttle_field = batchid." - ".jobid | ...
And then just choose that field for the throttling logic?