[Bug] Scheduling re-archiving of custom reports results in archives being deleted for unrelated reports

[snake14]

What happened?

When creating a new custom report, it schedules re-archiving so that the past n months (rearchive_reports_in_past_last_n_months) can be archived for the new report. Invalidation records are created for that specific report ID for the past n months. However, when archiving runs, it deletes the archives for other custom reports and doesn't archive the new report for those periods or re-archive the other reports.

What should happen?

The archives of existing reports should be untouched and the new report should be archived for the past n months, based on the rearchive_reports_in_past_last_n_months config.

How can this be reproduced?

First, I run the customreports:archive command to make sure that all of the archives have been created:

After creating a new custom report, the blob records are unchanged, but the re-archiving is scheduled:

When I run the archive process (core:archive) with a breakpoint before the API calls to archive:


I confirmed that all of the invalidation records had the same report ID.

Once archiving is done running:

More information is available in my notes on Jira tickets PG-2881 and PG-3173

Matomo major version

Matomo 5

Matomo minor or patch Version

5.0.0-rc8

PHP version

8.1.26

Server operating system

Ubuntu 22.04

What browsers are you seeing the problem on?

Not applicable (e.g. an API call etc.)

Computer operating system

Ubuntu 22.04

Relevant log output

No response

Validations

• Read our Contributing Guidelines.
• Follow our Security Policy.
• Check that there isn't already an issue that reports the same bug to avoid creating duplicates.
• The provided steps to reproduce is a minimal reproducible of the Bug.

[mneudert]

There seems to be something wrong with detecting the RecordBuilder:

matomo/core/ArchiveProcessor/RecordBuilder.php

Lines 328 to 330 in 1474562

	if (in_array($record->getName(), $requestedReports)) {
	return true;
	}

When running the archiving I received some values that obviously won't match:

$recordMetadata = {object[1]} 
 0 = {Piwik\ArchiveProcessor\Record} 
  type = "blob"
  name = "CustomReports_customreport_13_4"
  columnToSortByBeforeTruncation = "nb_visits"
  maxRowsInTable = null
  maxRowsInSubtable = null
  countOfRecordName = null
  countOfRecordNameIsRecursive = false
  columnToRenameAfterAggregation = null
  blobColumnAggregationOps = null
  multiplePeriodTransform = null
$requestedReports = {string[1]} ["13"]
 0 = "13"

If I do some hacky modifications to match that record the previously missing archive blobs appear as expected. Perhaps there is only a call to Archiver::makeRecordName to be patched into an override of RecordBuilder::isBuilderForAtLeastOneOf?

[michalkleiner]

Do you think it's an issue in CustomReports, @mneudert, after changing to use RecordBuilder?