Search performance issues after Splunk Upgrade 8.0.7 to 8.2.3 with mostly metrics based indexes

[walkerhound]

After we upgraded from 8.0.7 to 8.2.3, we are having lots of problems with search performance.  We noticed that the analytics workspace changed a great deal and we wonder if that could be causing the performance issue.

Now we have lots of searches queued - that didn't happen before.  Also sometimes the maximum number of historical searches is exceeded and we end up having to restart Splunk.  After the restart, our system runs okay for a while.

Any help will be appreciated.

[walkerhound]

I am told this problem is reflected in SPL-216787.  We have been advised to upgrade to 8.2.4 and change the parameter job_default_auto_cancel to 62.  I will update this question later to say if our problem has been solved.