planner: avoid exceeding the configured concurrency limit

[hawkingrei]

What problem does this PR solve?

Issue Number: close #61785

Problem Summary:

The issue is that customers have observed higher I/O consumption when the analyze operation reaches the index, compared to when it analyzes regular tables. (The analyze status contains sensitive information, so it will not be included here.)

The root cause of the issue lies in improper coding practices. When we perform the analyze operation, we create multiple concurrent tasks to execute it. However, within these concurrently spawned goroutines, we further create additional concurrency. This nested concurrency results in an actual level of parallelism that is significantly higher than we anticipated.

CREATE TABLE `test` (
  `c1` binary(16) NOT NULL,
  `c2` tinyint(1) NOT NULL DEFAULT '0',
  `c3` int NOT NULL,
  `c4` varchar(48) COLLATE utf8mb4_general_ci NOT NULL,
  `c5` varchar(512) COLLATE utf8mb4_general_ci DEFAULT NULL,
  `c6` enum('A','B','C') COLLATE utf8mb4_general_ci DEFAULT NULL,
  `c7` int unsigned NOT NULL DEFAULT '0',
  `c8` int unsigned NOT NULL DEFAULT '0',
  `c9` tinyint(1) GENERATED ALWAYS AS (`c7` > 0) VIRTUAL NOT NULL,
  `c10` int DEFAULT NULL,
  `c11` datetime(3) NOT NULL DEFAULT CURRENT_TIMESTAMP(3),
  `c12` datetime(3) NOT NULL DEFAULT CURRENT_TIMESTAMP(3),
  PRIMARY KEY (`c1`) /*T![clustered_index] CLUSTERED */,
  KEY `idx_c4_c2_c9_c3_c12_c5_c6` (`c4`,`c2`,`c9`,`c3`,`c12`,`c5`,`c6`),
  KEY `idx_c4_c2_c9_c12_c5_c6` (`c4`,`c2`,`c9`,`c12`,`c5`,`c6`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_general_ci;

analyze table chat_session all columns ;

show analyze status

+--------------+------------+----------------+-----------------------------------------------------------------------------------------------------------------+----------------+---------------------+---------------------+----------+-------------+----------------+------------+-------------------+----------+----------------------+
| Table_schema | Table_name | Partition_name | Job_info                                                                                                        | Processed_rows | Start_time          | End_time            | State    | Fail_reason | Instance       | Process_ID | Remaining_seconds | Progress | Estimated_total_rows |
+--------------+------------+----------------+-----------------------------------------------------------------------------------------------------------------+----------------+---------------------+---------------------+----------+-------------+----------------+------------+-------------------+----------+----------------------+
| test         | test       |                | analyze ndv for index idx_c4_c2_c9_c12_c5_c6                                                                    | 0              | 2025-06-18 14:48:05 | 2025-06-18 14:48:05 | finished | <null>      | 127.0.0.1:4000 | <null>     | <null>            | <null>   | <null>               |
| test         | test       |                | analyze ndv for index idx_c4_c2_c9_c3_c12_c5_c6                                                                 | 0              | 2025-06-18 14:48:05 | 2025-06-18 14:48:05 | finished | <null>      | 127.0.0.1:4000 | <null>     | <null>            | <null>   | <null>               |
| test         | test       |                | analyze table all indexes, columns c1, c2, c3, c4, c5, c6, c7, c9, c12 with 256 buckets, 100 topn, 1 samplerate | 0              | 2025-06-18 14:48:05 | 2025-06-18 14:48:05 | finished | <null>      | 127.0.0.1:4000 | <null>     | <null>            | <null>   | <null>               |
+--------------+------------+----------------+-----------------------------------------------------------------------------------------------------------------+----------------+---------------------+---------------------+----------+-------------+----------------+------------+-------------------+----------+----------------------+

You will see that it will create two task about analyze ndv for index.

the problem is here.

The first creation of concurrency

tidb/pkg/executor/analyze.go

Lines 121 to 126 in 8fc1430

	// Start workers with channel to collect results.
	taskCh := make(chan *analyzeTask, buildStatsConcurrency)
	resultsCh := make(chan *statistics.AnalyzeResults, 1)
	for range buildStatsConcurrency {
	e.wg.Run(func() { e.analyzeWorker(taskCh, resultsCh) })
	}

The second creation of concurrency

AnalyzeExec.analyzeWorker -> analyzeColumnsPushDownEntry -> analyzeColumnsPushDownV2

https://github.com/pingcap/tidb/blob/master/pkg/executor/analyze_col_v2.go#L105-L107

The third creation of concurrency

tidb/pkg/executor/analyze_col_v2.go

Lines 461 to 466 in 8fc1430

	var subIndexWorkerWg = NewAnalyzeResultsNotifyWaitGroupWrapper(resultsCh)
	subIndexWorkerWg.Add(statsConcurrncy)
	for range statsConcurrncy {
	subIndexWorkerWg.Run(func() { e.subIndexWorkerForNDV(taskCh, resultsCh) })
	}
	for _, task := range tasks {

This part is actually the most dangerous. It allows the concurrency of handleNDVForSpecialIndexes and the concurrency of column collection to coexist, which increases the business risk.

What changed and how does it work?

1、Wait untilhandleNDVForSpecialIndexesis completed before proceeding with the statistics collection for columns.

2、To prevent modifying the build stats concurrency, which could result in an exponential relationship in the actual number of concurrent tasks, we set the concurrency here to be the same as the build sampling concurrency.

Check List

Tests

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

  • I checked and no code files have been changed.

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

Please refer to Release Notes Language Style Guide to write a quality release note.

None

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 73.5170%. Comparing base (8fc1430) to head (a5f1a0d).
Report is 4 commits behind head on master.

Additional details and impacted files

@@               Coverage Diff                @@
##             master     #61786        +/-   ##
================================================
+ Coverage   73.0434%   73.5170%   +0.4735%     
================================================
  Files          1730       1730                
  Lines        481687     481990       +303     
================================================
+ Hits         351841     354345      +2504     
+ Misses       108289     106161      -2128     
+ Partials      21557      21484        -73     

Flag	Coverage Δ
integration	42.3912% <100.0000%> (?)
unit	72.3775% <100.0000%> (+0.0754%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Components	Coverage Δ
dumpling	52.7804% <ø> (ø)
parser	∅ <ø> (∅)
br	46.6118% <ø> (+0.0286%)	⬆️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[hawkingrei]

/retest

[hawkingrei]

/retest

[hawkingrei]

/retest

[ti-chi-bot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: AilinKid, winoros

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

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [AilinKid,winoros]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[ti-chi-bot]

[LGTM Timeline notifier]

Timeline:

• 2025-06-18 09:59:49.727862186 +0000 UTC m=+265842.451041165: ☑️ agreed by AilinKid.
• 2025-06-18 12:22:22.307701911 +0000 UTC m=+274395.030880890: ☑️ agreed by winoros.

[hawkingrei]

/retest

[hawkingrei]

/cherrypick release-8.5
/cherrypick release-8.1
/cherrypick release-7.5

[ti-chi-bot]

@hawkingrei: new pull request created to branch release-7.5: #61813.
But this PR has conflicts, please resolve them!

In response to this:

/cherrypick release-8.5
/cherrypick release-8.1
/cherrypick release-7.5

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 ti-community-infra/tichi repository.

[ti-chi-bot]

@hawkingrei: new pull request created to branch release-8.1: #61814.
But this PR has conflicts, please resolve them!

In response to this:

/cherrypick release-8.5
/cherrypick release-8.1
/cherrypick release-7.5

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 ti-community-infra/tichi repository.

[ti-chi-bot]

@hawkingrei: new pull request created to branch release-8.5: #61815.
But this PR has conflicts, please resolve them!

In response to this:

/cherrypick release-8.5
/cherrypick release-8.1
/cherrypick release-7.5

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 ti-community-infra/tichi repository.