Throttle openQA jobs resources based on test parameters size

[m-dati]

Define a priority component for some resources to control, to be added to a job priority setting at test start time, to limit the jobs resources requests.

Store the configuration for all the needed resources in openqa.ini file, in prio_throttling_parameters parameter, in a string of a format described in the ticket: "param1:scale1[:reference1],param2:scale2,...", to load the control values at openQA server start.

• Ticket poo#192952

[okurz]

In your commit message please write

Related progress issue: https://progress.opensuse.org/issues/192952

Please avoid the wording "malus" as it sounds so negative :) What we do is scale priority basically in both directions

[perlpunk]

it seems that https://app.circleci.com/pipelines/github/os-autoinst/openQA/19199/workflows/5a5b04f1-4a01-4606-8df0-4208907ae0b1/jobs/190404

Can't call method "config" on an undefined value at /home/squamata/project/lib/OpenQA/Schema/ResultSet/Jobs.pm line 170.
[19:01:53] t/ui/10-tests_overview.t ................... 
Dubious, test returned 255 (wstat 65280, 0xff00)
No subtests run 

is the same problem as the one we fixed in 21-needles.t.
In this case, just put a new Test::Mojo app before prepare_database:

my $t = Test::Mojo->new('OpenQA::WebAPI');
prepare_database;

and add it to the same commit as the 21-needles.t fix

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.56%. Comparing base (ce2072d) to head (3557cb4).
⚠️ Report is 5 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #6952   +/-   ##
=======================================
  Coverage   99.56%   99.56%           
=======================================
  Files         416      416           
  Lines       42396    42474   +78     
=======================================
+ Hits        42210    42288   +78     
  Misses        186      186           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 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.

[m-dati]

Going to apply the last discussed throttling logic, that includes a reference value, as described in ticket note-10

[okurz]

The changes look good. The still open points are something that we better solve in follow-up PRs. I only recommend to fix some issues in the git commit messages:

1. s/parmeters/parameters/
2. s/fix(test): Create/fix(test): create/
3. "[feat(throttling): refactor the configuration logic". Either "feat" or "refactor", can't be both :) IMHO the commit should be squashed with ef8373f and e6bd688 anyway

[m-dati]

About the non-squashed fix(test) commits, I did so after a peer-coding section, fixing unit tests impacted by the new code, having been suggested to keep separated the commits.
Therefore there are commit 'jumps', making the feat commits separated; willing to reorder the git hashes to unify those three in one, could cause conflicts in the code changes. But I try it.

[m-dati]

Really unclear what CI error is: despite All tests successful. Result: PASS, then not ok - unhandled output found

[okurz]

Really unclear what CI error is: despite All tests successful. Result: PASS, then not ok - unhandled output found

It says "not ok - unhandled output found" so you need to ensure that tests run without any output. There must only be the line "$testname .... ok", nothing else.

See

[2026-02-08 16:31:37.49638] [2279] [warn] No configuration files supplied, will fallback to default configuration
[16:31:38] t/config.t ................................................ ok     1728 ms ( 0.01 usr  0.00 sy…

[m-dati]

CI successful, finally.

[perlpunk]

Remember that split always splits on a regex, even if you pass it as a simple string.

[perlpunk]

We were in a collab session and refactored the throttling into its own function.
I pushed a commit on top.