Improve error handling for Lookup Table Cache purging

As continuation of https://github.com/Graylog2/graylog2-server/pull/10605 we could improve error handling on the FE side.

## Expected Behavior
The error message is expected to be enriched with details about nodes which failed to purge the cache and maybe with some further hints (e.g. regarding node availability)

## Current Behavior

## Steps to Reproduce (for bugs)

1. Locally: make the node endpoint (`LookupTableResource#performPurge`) unreachable, for instance by changing its path.
2. Set up a new Lookup Table
3. Try purging its cache

## Context
Original request: https://github.com/Graylog2/graylog2-server/issues/9554