Remove the Serverless label from ILM docs

[yetanothertw]

Fixes #1196

Removes the serverless applies_to tag from the ILM section of the Data lifecycle docs.

Preview available here

[eedugon]

LGTM!

One small comment, not really related with the scope of this PR.

In the Data Tiers page of Data lifecycle we have the "Serverless" tag, and I'm not sure if we should remove it from there too. Just for your consideration and analysis.

[yetanothertw]

Thanks for the review, @eedugon!

To address your comment, I think the Data tiers page applies to both (so both Serverless and Elastic Stack labels apply) judging by this snippet:

You can automate these lifecycle transitions using the data stream lifecycle, or custom index lifecycle management.

Just to doublecheck, @shainaraskas or @kilfoyle , could either of you confirm, please?

Many thanks,
Vlada

[shainaraskas]

@yetanothertw @eedugon is correct that data tiers are generally a stack concept only.

data tiers specifically are meant to allow you to configure different hardware for different categories of data (very simplified definition).

as a rule, anything that talks about hardware/sizing centric concepts (nodes, shards, hot spotting all fall into this category) does not apply to serverless. this is because the value prop of serverless is that you don't have to think about hardware or sizing at all.

in stack, technically data stream lifecycle might interact with tiers just by virtue of running on hardware that is set up to use tiers, but it doesn't do a great job of interacting with tiers, especially granular ones. the callout on the tiers page might be a little misleading. but data stream lifecycle doesn't rely on manual configuration of tiers, making DSL serverless-compatible.

this page is probably our most accurate in terms of "what applies to serverless" - when in doubt, you can check there as well. it touches on tiers.

ALL THAT BEING SAID: serverlessshould be removed from the data tiers page. if you want to go the extra mile, the page also deserves a note explaining why you don't have to care about tiers for serverless, similar to this one on the stack settings page. helps people to understand that it's not REMOVED functionality, but instead AUTOMATED in a way that makes cluster management easy.

will take a look at the rest of the pr shortly - sorry for wall of text :)

[shainaraskas]

similarly, I don't think serverless supports curator. not sure about that one.

[shainaraskas]

ah I thought I submitted a review but I didn't - these tag removals look good. see my other comments re: data tiers and curator

[yetanothertw]

Thank you for detailed response @shainaraskas -- I've removed the labels from curator and data tiers.

I want to add the admonition you suggested to make it clear that Serverless manages data tier concepts for you. Does this look right to you: https://docs-v3-preview.elastic.dev/elastic/docs-content/pull/1531/manage-data/lifecycle/data-tiers

I wasn't sure if I should keep the link to project settings, like in the admonition you referenced, as it doesn't seem relevant.

[yetanothertw]

Opened issue #1547 to address the admonition suggestion separately. Thanks all.