Page MenuHomePhabricator

Maniphest Reports break when you have 75k tasks
Closed, ResolvedPublic


Upstream task:

This was a problem that we saw during the Bugzilla migration weekend. Reporting it here. We should report this upstream. I think Chase or Mukunda had some IRC chat with the maintainers, but no task submitted?

If you go to you get

Request aborted by debug time limit after 30 seconds.

ManiphestTask.php:153 PhabricatorStartup->onDebugTick()
ManiphestTaskQuery.php:349 ManiphestTask->attachProjectPHIDs()
PhabricatorPolicyAwareQuery.php:260 ManiphestTaskQuery->didFilterPage()
ManiphestReportController.php:409 PhabricatorPolicyAwareQuery->execute()
ManiphestReportController.php:46 ManiphestReportController->renderOpenTasks()
AphrontController.php:33 ManiphestReportController->processRequest()
index.php:103 AphrontController->handleRequest()

Related Objects

View Standalone Graph
This task is connected to more than 200 other tasks. Only direct parents and subtasks are shown here. Use View Standalone Graph to show more of the graph.

Event Timeline

Qgil raised the priority of this task from to Medium.
Qgil updated the task description. (Show Details)
Qgil changed Security from none to None.
Qgil subscribed.

@mmodell put that in but not sure if there's a corresponding upstream ticket yet?

Qgil lowered the priority of this task from Medium to Lowest.Dec 2 2014, 9:06 AM

Evan's reply upstream: (Building reporting and data systems) is the real solution to this (...)

Alright, this means a total rewrite. And this means that we are not going to have a short term solution unless someone finds a hackish solution to the current code. There is some discussion in

Meanwhile, I will set realistic priority here. For short term Wikimedia Phabricator reports, see T1003: Monthly report of total / active Phabricator users.

It's probably not a good idea to have a page that hangs for 30 seconds displayed so prominently. The link should be removed/hidden from the maniphest page until a solution is found.

Patch landed upstream:

With the patch, the reports page generates in ~500ms now on this install with 6.5K tasks, so I'd expect it to take ~6s to generate on WMF with 75K tasks if everything scales linearly (which it hopefully should). Still not great, but way better than totally broken. Let me know if you actually see that sort of runtime?

Great catch on the getDestinationPHIDs() issue, in any case.

Qgil claimed this task.

This was fixed after our last upgrade. Now loading takes some seconds, but it loads reliably.