How to re-run a scheduled correlation search after planned downtime?

[sheamus69]

Where there is a planned scheduled outage of a network device, which will effectively kill many of the feeds due to come into splunk, what would be the best approach to re-running correlation searches on the missing feeds once connection has been restored?

[kamal_jagga · Accepted solution]

You can use the curl command to trigger the searches from back-end without changing its properties.
eg. curl -k -u username:Password https://localhost:9086/services/saved/searches/testkamal2/reschedule -X POST

This will immediately trigger your search. You can write a script to trigger the queries one by one every few mins.

[sheamus69]

How would this work for effectively looking at the missed data from the restored feed? Surely that command would only run the search now, not step back through the data? Or am I missing something obvious (admittedly, highly likely)?

[FrankVl]

Aren't correlation searches real time searches that run based on index time, rather than event time?

[sheamus69]

Nope, they run at scheduled intervals, and if the event feed is delayed, things can be missed.

[sheamus69]

Just did a little reading into this approach, and I think you are right, as shown in the example from the rest api page:

curl -k -u admin:pass https://localhost:8089/services/saved/searches/Purchased%20products%2C%20last%2024%20hours/reschedul... -d schedule_time=2012-08-15T14:11:01Z