Hi all,
We are using SoS app for monitoring our schedules. We are working on reducing the schedules which are being skipped . But what is case about the deferred searches ? Its keep on increasing ? how to reduce them.
thoughts pls
A deferred search is a search that couldn't be executed right now, because of the system or role search concurrency limit. Therefore they are executed a few seconds later. This is an expected behavior.
By example is you have a dashboard with 10 searches, but a limit of 6 concurrent searches, some panels will load after the first ones completed.
If you look at the audit logs, you can find how long they were deferred before being executed.
If a search is deferred too long, it will finally be skipped: skipped searches
The root cause are usually caused by :
- too many searches : you have too many searches (or heavy dashboard)
- non optimized scheduled searches taking long to run and overlapping.
- hardware limit : the indexers and search-head have not enough cpu core to handle high search concurrency. (check limits.conf), on 6.0 the formula for historical search concurrency = 6+ 1* (number of cores)
(hint, disable the deployment monitor app if you have it)
Thanks a lot
more an advice then an answer, check out this answer on search scheduling http://answers.splunk.com/answers/33717/scheduled-searches-for-summary-index-does-not-run-no-skipped...