Bug 422227

Summary:	org.kde.plasmashell Internal Server Error
Product:	[Frameworks and Libraries] frameworks-kuserfeedback	Reporter:	Andres Betts <anditosan1000>
Component:	Management and Analytics Console	Assignee:	Volker Krause <vkrause>
Status:	RESOLVED FIXED
Severity:	normal	CC:	kde
Priority:	NOR
Version First Reported In:	unspecified
Target Milestone:	---
Platform:	Other
OS:	Linux
Latest Commit:	https://invent.kde.org/libraries/kuserfeedback/commit/f22b81e8273ddbd79ff249f2f06bbc87bf5074d4	Version Fixed/Implemented In:
Sentry Crash Report:

Description Andres Betts 2020-05-29 13:49:12 UTC

SUMMARY

Hi,

The console shows 

Error transferring https://telemetry.kde.org/analytics/data/org.kde.plasmashell - server replied: Internal Server Error

When trying to access org.kde.plasmashell

Is there anything we can do about it?

STEPS TO REPRODUCE
1. 
2. 
3. 

OBSERVED RESULT


EXPECTED RESULT


SOFTWARE/OS VERSIONS
Windows: 
macOS: 
Linux/KDE Plasma: 
(available in About System)
KDE Plasma Version: 
KDE Frameworks Version: 
Qt Version: 

ADDITIONAL INFORMATION

Comment 1 Volker Krause 2020-06-02 06:10:50 UTC

*** Bug 422348 has been marked as a duplicate of this bug. ***

Comment 2 Volker Krause 2020-06-02 06:12:15 UTC

Plasma and Discover show this problem, all other products seems to work fine. Asking sysadmins for the server logs.

Comment 3 Volker Krause 2020-06-02 06:56:04 UTC

Ben checked the logs, the problem seems to be that the server hits the PHP memory limit for those two products (as they have accumulated 10+M of data by now), and right now the full JSON response with that data is assembled in memory.

Comment 4 Bug Janitor Service 2020-08-21 08:42:35 UTC

A possibly relevant merge request was started @ https://invent.kde.org/libraries/kuserfeedback/-/merge_requests/4

Comment 5 David Edmundson 2020-08-21 15:54:20 UTC

Git commit f22b81e8273ddbd79ff249f2f06bbc87bf5074d4 by David Edmundson.
Committed on 21/08/2020 at 08:42.
Pushed by davidedmundson into branch 'master'.

[server] Convert JSON fetch to stream data

Currently the code builds up all the data in a massive blob then tries
to encode the JSON. This hits a memory limit on our server for popular
products.

By fetching and printing one piece of information at a time we should
avoid this.  Fetching scalar and non-scalar information now needed to be
put together so we fetch non-scalar information on a per-row basis.

Technically more queries, but it saves hash look ups our side.

Tested against the submittest unit test. I haven't tested with "real"
data.

M  +1    -1    src/server/shared/analytics.php
M  +40   -44   src/server/shared/sample.php

https://invent.kde.org/libraries/kuserfeedback/commit/f22b81e8273ddbd79ff249f2f06bbc87bf5074d4