Restore snapshot logs include too many indices

[DaveCTurner]

The log message added in #88257 to indicate the start of a snapshot restore lists all the indices in the snapshot rather than just the ones we are restoring, which is confusing to users. Also the log message at the end doesn't indicate to which restore it pertains, and a user may issue in parallel several restore requests for indices from the same snapshot.

I think we should:

• restrict the list of indices in the started restore of snapshot [... message to mention just the indices being restored
• consider also calling out the feature states and whether we're restoring the global state
• add the restore UUID to both messages, i.e. started restore [--UUID--] of snapshot [... and completed restore [--UUID--] of snapshot [... so that we can correlate the starts and ends.

[elasticsearchmachine]

Pinging @elastic/es-distributed (Team:Distributed)