This is an automated cherry-pick of #18439

What is changed and how it works?

Issue Number: Close #18463

What's Changed:

In previous work #17801, TiKV has introduced the detection mechanism for kvdb disk to detect I/O hang issues.

However, recent customer feedback highlighted the need to extend detection coverage to I/O jitters, ensuring TiKV can automatically recover from abnormal states caused by KVDB I/O jitters.

Therefore, this ticket is built to tracks the development efforts to enhance TiKV’s I/O jitter detection and recovery mechanism. And the majority parts of this change are listed as followings show:

• for configurations:
  • raftstore.inspect_kvdb_interval: 2s -> 100ms
• for detection mechanism on kvdb:
  • SlowScore::ratio_thresh：60% -> 10%

Enhances the detection mechanism to cover the I/O jitters on kvdb disk if deploys with separated mount paths.

Related changes

• PR to update pingcap/docs/pingcap/docs-cn:
• Need to cherry-pick to the release branch

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
• No code

Workloads	v8.5.1	With this PR
Special workloads
tpcc 1k warehouses
Sysbench - oltp_read_write

Side effects

• Performance regression: Consumes more CPU
• Performance regression: Consumes more Memory
• Breaking backward compatibility

Release note

Enhances the detection mechanism to cover the I/O jitters on kvdb disk if deploys with separated mount paths.