Optimize Index Management's indices list network performance

[cjcenizal]

When Index Management is loaded, the UI sends a request to GET /api/index_management/indices. The Kibana server handles this request by making two requests to Elasticsearch:

• GET *?expand_wildcards=hidden,all
• GET /_cat/indices?format=json&h=health,status,index,uuid,pri,rep,docs.count,sth,store.size&expand_wildcards=hidden,all&index=*

However, the table itself only uses information from the /_cat/indices request. The request for all indices contains settings, aliases, and mappings information, none of which is surfaced in the table:

This means that the initial request to load all indices can potentially be very slow or even crash Kibana. I opened elastic/elasticsearch#64099 to address this problem through improved pagination of Elasticsearch REST APIs, but we can solve this specific problem in Kibana by implementing the following:

• Request only the /_cat/indices API when listing all indices.
• When fetching an individual index, e.g. when opening a flyout to show index details, request the additional information provided by the Index API.

[elasticmachine]

Pinging @elastic/kibana-stack-management (Team:Stack Management)

[cjcenizal]

Did some digging at the cat indices API currently lacks the following information that we need to implement the current UX:

• hidden - Whether the index is hidden or not
• data_stream - The name of the data stream that owns the index, if there is one
• aliases - The aliases that point to this index (implemented in the API, but not consumed by the UI, possibly an oversight)