Batch add-block-index-to-close, add-index-block and finalize-index-block tasks

[DaveCTurner]

The process of closing an index involves multiple cluster state updates to add and finalise its blocks; there are also other reasons for adding blocks to an index. ILM may close an index to adjust its codec prior to force-merging, and may apply a write block for other reasons. Today the associated cluster state update tasks run at priority URGENT and are not batched, so can take a very long time to process and gum up other things.

I encountered a cluster that was processing a few thousand such tasks at a rate of about 20 per minute.

We should batch these tasks together to avoid this causing a problem in future.

Relates #77466

[elasticmachine]

Pinging @elastic/es-data-management (Team:Data Management)

[joegallo]

Handled by way of #83760, #84259, and #84374.