Optimize apply speed under heavy write pressure (#4883)

[ti-chi-bot]

This is an automated cherry-pick of #4883

What problem does this PR solve?

Issue Number: ref #4728

Problem Summary: When tiflash is under heavy write pressure, it will consume a lot of write throughput and cause some latency on the process of apply raft log, which will make tiflash consume a log memory.

What is changed and how it works?

1. Only allow one delta flush process on a segment to process at any time to avoid unnecessary throughput consumption;
2. When processing CompactLog raft command, just try to flush the unsaved value in storage once and return Persist if the flush succed or None if the flush failed.(Originally we try to flush until it success)
3. Increase foreground flush threshold for write thread.
4. Avoid add minor compaction task to background if the segment is flushing to save some throughput and reduce lock contention on the segment.

Check List

Tests

• Manual test (add detailed scripts or steps below)
  Test step:

1. perform large transaction update through 5 tidb instance at the same time and the total transaction size of all tidb instance is about 24.5 GB;
   Before optimize:
   
   After optimize:
   

Detail result(unit: GB)

	1st run	2nd run	3rd run	4th run	5th run	6th run)	7th run	average value	amplification factor
before optimization	40.8	37.2	37.3	36	34.7	33.6	34.5	36.3	1.48
after optimization	38.1	29.4	29	34.3	29.6	32.6	37.7	32.9	1.34

Side effects

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

Documentation

• Affects user behaviors
• Contains syntax changes
• Contains variable changes
• Contains experimental features
• Changes MySQL compatibility

Release note

None

[ti-chi-bot]

[REVIEW NOTIFICATION]

This pull request has been approved by:

• CalvinNeo

To complete the pull request process, please ask the reviewers in the list to review by filling /cc @reviewer in the comment.
After your PR has acquired the required number of LGTMs, you can assign this pull request to the committer in the list by filling /assign @committer in the comment to help you merge this pull request.

The full list of commands accepted by this bot can be found here.

Reviewer can indicate their review by submitting an approval review.
Reviewer can cancel approval by submitting a request changes review.

[ti-chi-bot]

@ti-chi-bot: This cherry pick PR is for a release branch and has not yet been approved by release team.
Adding the do-not-merge/cherry-pick-not-approved label.

To merge this cherry pick, it must first be approved by the collaborators.

AFTER it has been approved by collaborators, please ping the release team in a comment to request a cherry pick review.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.