Sourcetype and indexes

[kam_emea]

Hi

I am trying to understand how indexes and sourcetype are defined. Let's say I have an app with a web component and a database component.

Should the web component and db component be different indexes? 

And the sourcetype is a category within each index? Does Splunk automatically determine the sourcetype based on the data it ingested? Or is this something that is manually done?

[richgalloway · Accepted solution]

An index is a repository for data.  A sourcetype describes a kind of data.  The web component and database component almost certainly are different sourcetypes, but they don't necessarily need separate indexes.

Create a new index when the data either: 1) has different retention requirements from other data; or 2) has different access requirements from other data.  There are performance considerations, too, but most sites don't have to worry about them. 

Data that often will be accessed in the same search can be put in the same index for faster searching.

Yes, Splunk can automatically determine the sourcetype of your data, but it can get it wrong so one should always define sourcetypes explicitly in props.conf.

[isoutamo]

Hi
one old answer to this with links to two conf presentations. Those are still valid even those are for quite old splunk versions.
https://community.splunk.com/t5/Splunk-Search/Index-vs-Sourcetype-What-s-faster/m-p/468365
r. Ismo