Do data streams require rollover_alias setting

Whilst setting up a new heartbeat-8.3.2 I've noticed that the automatically created template has only index.lifecycle.name setting but not the rollover_alias. Does that mean it's not required when using data streams? The rollover process seems to be working although not at frequency set e.g. I set ILM to rollover every 1min (just for testing purposes) but the index didn't rollover for another 5min+ minutes despite new data continuously coming in.

1 Like

That's because the ILM processes run in the background on a polling sequence. It will never run at one minute unless you tweak a number of settings.

This is pretty common experience when people set up test. Expect things to roll over and megabytes or minutes, but keep in mind. ILM meant to be used at gigabyte hours or days.

indices.lifecycle.poll_interval
(Dynamic, time unit value) How often index lifecycle management checks for indices that meet policy criteria. Defaults to 10m

Thanks for the explanation. As you've stated we wouldn't set ILM to 1m in Production anyway but thank you for the extra information. What about the question re rollover_alias setting? Is it not required with data streams? Thanks

No write_alias / rollover_alias is Not required it is handled behind the stream. You can think of the data stream name as the "entry point" then all the actual concrete indices are managed behind the scenes

1 Like

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.