Cassandra ReadTimeout when cooking a git bare bundle with the vault

This type of error is happening quite a lot since recently.

I can also reproduce locally.

For the record, vault cooking is fetching directories and contents in parallel using multi-threading.

Sentry Issue: SWH-VAULT-8W

RemoteException: <RemoteException 500 ReadTimeout: ['Error from server: code=1200 [Coordinator node timed out waiting for replica nodes\' responses] message="Operation timed out - received only 0 responses." info={\'consistency\': \'LOCAL_QUORUM\', \'required_responses\': 2, \'received_responses\': 0}']>
(13 additional frame(s) were not displayed)
...
  File "/opt/swh/.local/lib/python3.10/site-packages/swh/core/api/__init__.py", line 196, in meth_
    return self._post(meth._endpoint_path, post_data)
  File "/opt/swh/.local/lib/python3.10/site-packages/swh/core/api/__init__.py", line 356, in _post
    return self._decode_response(response)
  File "/opt/swh/.local/lib/python3.10/site-packages/swh/core/api/__init__.py", line 443, in _decode_response
    self.raise_for_status(response)
  File "/opt/swh/.local/lib/python3.10/site-packages/swh/storage/api/client.py", line 45, in raise_for_status
    super().raise_for_status(response)
  File "/opt/swh/.local/lib/python3.10/site-packages/swh/core/api/__init__.py", line 433, in raise_for_status
    raise exception from None

Bundle cooking failed.

It's possibly a duplicate of [1]

Common denominator being cassandra time-outing during an operation started by either the storage [1] or the vault (here).

[1] #5444