How to maintain lookups?

[sidhantbhayana]

I am analyzing our Splunk set-up and was going through the lookups, need suggestions on the best strategy to maintain lookups.
1. bundle size
2. switching over to manual lookups(instead of automatic lookups)
3. Performance issues with automatic lookups

Would appreciate if something else can be analyzed or taken care of.

Thanks,
SB

[woodcock · Accepted solution]

Splunk will let you know when the bundle size is too big, don't worry about that. I use automatic lookups whenever they make sense; when it is useful for everybody. If the data has a code and you can translate to a description (like HTTP response codes), then I set it up as automatic. If it is only for me, I just do it in my search manually. The only performance issues you may have is if you use the hacky apply this to all sourcetypes thing.

[sidhantbhayana]

Thanks @woodcock, agree to your response! I am planning to blacklist few lookups as the bundle size has reached ~1.5GB and it kind of becomes an issue in search peers.
Also, currently evaluating replacing lookups with kvstore, got to know that Splunk creates a csv from the kvstore depending on the query and pushes the csv as part of bundle.

Looking forward to understand how kvstore will work and would it be better to replace lookups, considering the fact that the data is incremental.

Thanks,
SB