refactor(query): mitigate exchange coordinator lock contention

[dqhl76]

I hereby agree to the terms of the CLA available at: https://docs.databend.com/dev/policies/cla/

Summary

refactor(query): mitigate exchange coordinator lock contention

When query concurrency is high, we observe frequent warnings caused by waiting too long to acquire the queries_coordinator lock in on_finished_query.

Waited 2.104701739s to acquire queries_coordinator lock

We try to refactor queries coordinator:

1. Split the single coarse-grained lock in the original queries coordinator into per-query independent locks to reduce lock contention.
2. Shorten the lock holding time, eliminate reentrancy, and replace the reentrant lock with a regular mutex.

Tests

• Unit Test
• Logic Test
• Benchmark Test
• No Test - Explain why

Type of change

• Bug Fix (non-breaking change which fixes an issue)
• New Feature (non-breaking change which adds functionality)
• Breaking Change (fix or feature that could cause existing functionality not to work as expected)
• Documentation Update
• Refactoring
• Performance Improvement
• Other (please describe):

---

This change is