Question regarding summary index with saved search

[chinmayc469]

Hello,

I have created a saved search to populate summary index. I am running saved search for every 5 minutes.

What i want is, first time when the saved search runs, it should run with time range as all time.
And from the second time on wards, saved search should with time range as "last 5 mins" (ie, latest=now and earliest=last time when ss ran succesfully)

So that i will avoid duplicate of data in summary index.

How to achieve this?

Thanks in advance.

[skoelpin]

Yep, run the first saved search manually then schedule it thereafter

[chinmayc469]

how about in production machines? we will not have access to change once it is gone to production.

[skoelpin]

I don't get what you're asking.. On your production servers, you should backfill against all-time then once your backfilled, you should set up a scheduled search which will populate your summary index going forward. I would also suggest using the python script to backfill. This gives you the ability to ignore data thats already been backfilled and you can run parallel backfills without duplicates

https://docs.splunk.com/Documentation/Splunk/7.1.2/Knowledge/Managesummaryindexgapsandoverlaps

[pradeepkumarg]

The first all time search, you can probably run manually, there after, you should chose earliest and latest with little bit of padding. latest=now is not a good practice as you are not accounting for the delay in the raw data. You can do something like earliest=-8m@m latest=-3m@m so that you account for 180 seconds of delay depending on how busy your indexers and forwarders are.