I am attempting to identify when Splunk users are running searches against historic data (over 180 days old). Additionally, as part of the same request, looking to identify where users have recovered data from DDAA to DDAS to run searches against that.
This is to build a greater understanding of how often historic data is accessed to help guide data retention requirements in Splunk Cloud (i.e. is retention set appropriately or can we extend/reduce retention periods based on the frequency of data access).
Hi
I have this query too find queries which search span is more than 90d.
index=_audit action=search info=completed NOT is_realtime=1 earliest=0
| eval search_et = if(search_et="N/A", 0, search_et)
| eval search_lt = if(search_lt="N/A", exec_time, search_lt)
| eval srch_window = ((search_lt-search_et)/86400)
| eval lookback = case( round(srch_window) <= 1, "-1d", round(srch_window) > 1 AND round(srch_window) <= 7, "1-7d", round(srch_window) > 7 AND round(srch_window) <= 10, "7-10d", round(srch_window) > 10 AND round(srch_window) <= 30, "11-30d", round(srch_window) > 30 AND round(srch_window) <= 60,"30-60d", round(srch_window) > 60 AND round(srch_window) <= 90, "60-90d", 1=1, "+90d" )
| search lookback="+90d"
| table user info event_count result_count search
| stats count avg(event_count) as avg_event avg(result_count) as avg_results values(info) as info by search, user
| sort 0 -count
You probably could modify it for your needs?
r. Ismo