[REVIEW NOTIFICATION]

This pull request has not been approved.

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.

Skipping CI for Draft Pull Request.
If you want CI signal for your change, please convert it to an actual PR.
You can still manually trigger a test run with /test all

@glorv @CabinfeverB @nolouch PTAL, thx!

Why need change them?

• The previous test data due to non-exclusive hard disk, io bandwidth is not full, so the tikv is not the bottleneck, the bottleneck is tidb.
  But because tidb is easy to expand, tikv is more difficult to expand, so generally tikv is easy to bottleneck.
  Now the test data are basically playing full tikv, may be slightly more accurate.
• cpu formula is ms, the previous calculation forgot to multiply by 1000. @nolouch

/review default

@HuSharp:

Pull Request Summary

Title: Resource control: update static calibrate params

PR Description:

This PR aims to solve issue #43212. The key changes are the updates in calibration constants for several workloads. The new test data now generally plays full tikv and may be slightly more accurate. This also fixes a calculation error in the CPU formula, which was missing the factor of 1000 to obtain ms. Some formatting updates have been made along with a few comments.

Tests:

• Unit test

Release Note:

- The previous test data due to non-exclusive hard disk, IO bandwidth is not full, so the tikv is not the bottleneck, the bottleneck is tidb.
But because tidb is easy to expand, tikv is more difficult to expand, so generally tikv is easy to bottleneck.
Now the test data are basically playing full tikv, may be slightly more accurate.
- CPU formula is ms, the previous calculation forgot to multiply by 1000.

Diff

+ Various calibration constant updates
+ CPU now multiplied by 1000 to convert into ms
+ Fixed comment typo and formatting issues

Review

The PR updates calibration parameters for several workloads, making them more accurate as a result of playing full tikv. Additionally, the CPU formula calculation was fixed by multiplying it with 1000 to obtain the correct unit. Formatting issues and comment typos were also corrected. Overall, the changes seem appropriate and should improve the performance monitoring and calibration procedure for TiDB.

Potential Problems

1. The updated calibration constants were taken from new test data, and it is mentioned they are more accurate, but it's not clear how these values were obtained and whether they have been rigorously tested.

Suggestions

1. Provide a detailed explanation of how the new calibration constants were obtained and if they have been thoroughly tested to validate their accuracy.

PR needs rebase.

Codecov Report

Merging #43541 (eef0191) into master (649fc4c) will increase coverage by 0.0071%.
Report is 1967 commits behind head on master.
The diff coverage is 100.0000%.

@@               Coverage Diff                @@
##             master     #43541        +/-   ##
================================================
+ Coverage   73.4725%   73.4796%   +0.0071%     
================================================
  Files          1190       1190                
  Lines        372738     373419       +681     
================================================
+ Hits         273860     274387       +527     
- Misses        81273      81446       +173     
+ Partials      17605      17586        -19     

@HuSharp: The following test failed, say /retest to rerun all failed tests or /retest-required to rerun all mandatory failed tests:

Full PR test history. Your PR dashboard.

maybe stale data