I recently upgraded my search head cluster to 9.x and since then my skipped/deferred searches have sky rocketed.
index=_internal source=*scheduler.log
status=* | timechart span=60s count by status
Please see this knowledge article for details on the known issue and fix/workaround:
https://splunk.my.site.com/customer/s/article/High-Skipped-Search-Ratio-after-upgrading-to-9-x