[fix][broker] BacklogMessageAge is not reset when cursor mdPosition is on an open ledger

[shibd]

Motivation

The topic stats have a metric: oldestBacklogMessageAgeSeconds, which represents the age in seconds of the earliest backlogged message's AgeSeconds in the topic.

The broker will periodically check each topic to update these metics, and it decides whether to use a precise check or an estimated check based on the PreciseTimeBasedBacklogQuotaCheck configuration.

When PreciseTimeBasedBacklogQuotaCheck=false, the broker will calculate its age based on the close time of the ledger corresponding to the old position.

Because the time can only be obtained when the ledger is closed, if an oldPosition was obtained initially, but afterward, the subscription in the topic is always caught up to the currently open ledger, then it will be skipped every time this oldPosition is updated. (The broker's default interval for updating the oldPosition is 60S, and it is very likely to hit this situation on every check.)

pulsar/pulsar-broker/src/main/java/org/apache/pulsar/broker/service/persistent/PersistentTopic.java

Lines 3846 to 3855 in 437d9f6

	EstimateTimeBasedBacklogQuotaCheckResult checkResult =
	estimatedTimeBasedBacklogQuotaCheck(oldestMarkDeletePosition);
	if (checkResult.getEstimatedOldestUnacknowledgedMessageTimestamp() != null) {
	updateResultIfNewer(
	new OldestPositionInfo(
	oldestMarkDeleteCursorInfo.getPosition(),
	oldestMarkDeleteCursorInfo.getCursor().getName(),
	checkResult.getEstimatedOldestUnacknowledgedMessageTimestamp(),
	oldestMarkDeleteCursorInfo.getVersion()));
	}

This can be very confusing for users using this metric, as the subscription's MarkDeletePosition is at the latest position, but the oldestBacklogMessageAgeSeconds metric keeps increasing.

You can run the backlogsAgeMetricsNoPreciseWithoutBacklogQuota test in this PR to reproduce this situation.

Modifications

When PreciseTimeBasedBacklogQuotaCheck is false, if the oldPosition is on a ledger that is currently open, we should consider that there is no oldPosition at this time and set it to -1.

Verifying this change

• Added backlogsAgeMetricsNoPreciseWithoutBacklogQuota to cover it.

Does this pull request potentially affect one of the following parts:

If the box was checked, please highlight the changes

• Dependencies (add or upgrade a dependency)
• The public API
• The schema
• The default values of configurations
• The threading model
• The binary protocol
• The REST endpoints
• The admin CLI options
• The metrics
• Anything that affects deployment

Documentation

• doc
• doc-required
• doc-not-needed
• doc-complete

Matching PR in forked repository

PR in forked repository:

[codecov-commenter]

Codecov Report

❌ Patch coverage is 0% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 74.26%. Comparing base (5e59d0e) to head (990851f).
⚠️ Report is 42 commits behind head on master.

Files with missing lines	Patch %	Lines
...sar/broker/service/persistent/PersistentTopic.java	0.00%	1 Missing ⚠️

Additional details and impacted files

@@              Coverage Diff              @@
##             master   #24915       +/-   ##
=============================================
+ Coverage     38.45%   74.26%   +35.81%     
- Complexity    13176    33844    +20668     
=============================================
  Files          1855     1913       +58     
  Lines        144926   149329     +4403     
  Branches      16814    17334      +520     
=============================================
+ Hits          55735   110905    +55170     
+ Misses        81668    29574    -52094     
- Partials       7523     8850     +1327     

Flag	Coverage Δ
inttests	26.42% <0.00%> (+0.03%)	⬆️
systests	22.83% <0.00%> (+0.07%)	⬆️
unittests	73.79% <0.00%> (+39.16%)	⬆️

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

Files with missing lines	Coverage Δ
...sar/broker/service/persistent/PersistentTopic.java	79.11% <0.00%> (+26.33%)	⬆️

... and 1414 files with indirect coverage changes

🚀 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.

[lhotari]

LGTM