Expose additional index attributes via _cat/indices

[jakelandis]

Some newer features have been introduce that would be good additions to the responses of the cat index API.

• Cross cluster replication - is an index a leader or follower ?
• Index Lifecycle management - is the index managed, if so, what phase is it in ?
• Frozen indicies - is the index frozen ? Note - this can be determined by the search.throttled attribute, but it would be nice if the attribute was specific to an index that has been frozen.
• Rollups - is the index a result of a roll up ?

All (or most) of this information can be found via other APIs, but index centric information would be nice to be exposed via the _cat/indices API.

[elasticmachine]

Pinging @elastic/es-core-features

[dakrone]

Marking this as team discuss because we're not sure whether these are exposed to the cat APIs, since they are in the open source code and these features (ILM and CCR in particular) are in x-pack

[bmcconaghy]

_cat/indices has info about frozen indices, and that is an xpack feature, so there is precedent for this.

[dakrone]

@bmcconaghy that was a previous implementation of frozen indices, because the engine was part of the core code, the current frozen indices implementation does not have any info about frozen status

EDIT: they do have status about it with the search.throttled property, but again, this is a property that is part of the core code, not x-pack

[jakelandis]

I have had a couple conversations on this topic. In general the desire here is for easy to consume API to surface this data. This may come in the form of when creating new indexes, paying additional attention the ease of consumption, or possible workflow specific (in this case UI workflow) API's that may be a subset of the actual APIs.

Using the cat API is one means to allow easier access to this type of data. I will remove the team-discuss label for now, but encourage when building out new API's to think of the expected workflows (like UI) and ensure that API is easily consumable.

We haven't made a decision on if using the cat API is correct place to make this data easier to consume, and leaving this issue open for further discussion.

[kurtado]

I was about to open an issue, but this one captures my thoughts pretty well, so I'll add my +1 here :)

It's certainly possible to get all the info about an index from mutiple APIs, but adding all/some of these to _cat would be super helpful. In particular, I was trying to get the ILM info, but it's not in _cat, and so needed to make additional API calls. I get that there may be x-pack/license issues...