Handle config updates in scenario where bucket is not found

There is a bug report about seeing bucket not_found responses like this

[2024-02-06 22:06:49.242] [1,29] [debug] 0ms, [a70699-c527-6c41-4a7c-263fd14b99ee54/a38b8b-3237-db4f-c164-4530d9c03ab436/tls/BUCKET_NAME] <IP_ADDRESS/IP_ADDRESS:11207> kv_engine node does not have configuration propagated yet (opcode=select_bucket (0x89), status=not_found (0x01), opaque=5)

Part of the problem has been fixed in CXXCBC-450, in particular when the session might still process event using old bootstrap handler and change the its state, while the re-bootstrap process is ongoing, but the new bootstrap handler has not been yet assigned.

This ticket has been opened as a tombstone, as it was hard to reproduce the original issue, and the work will continue once we will get more information.