Move wait for log uploads logic out of logger and tracer into pipeline runtime

[6543]

fix https://ci.woodpecker-ci.org/repos/3780/pipeline/33635/39 as the dummy backend just returns to fast and so sometimes the uploads wait group was just to slow, now we remove the wait from the logger and tracer implementation into the runtime where it belongs

PS: this race also ocures in real world usage, where the agent operates on ram disks

this is also the successor to #6263

[6543]

we get rid of all the hacks more and more 🎉

[codecov]

Codecov Report

❌ Patch coverage is 90.32258% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 41.19%. Comparing base (8e6084c) to head (7817902).

Files with missing lines	Patch %	Lines
pipeline/runtime/step.go	80.00%	2 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #6471      +/-   ##
==========================================
- Coverage   41.25%   41.19%   -0.06%     
==========================================
  Files         431      431              
  Lines       28792    28779      -13     
==========================================
- Hits        11878    11856      -22     
- Misses      15847    15853       +6     
- Partials     1067     1070       +3     

☔ 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.

[6543]

this is also the successor to #6263

[6543]

PS: if you want to verify non flakyness, just runn this on main and this pull head:

go test -count=300 -race -tags test -coverprofile=coverage.out ./e2e/...
go test -count=300 -race -tags test -coverprofile=coverage.out ./pipeline/runtime/...

(run can take a few minutes...(depending on how insane you go with count))

[6543]

current flaky-ness:

• https://ci.woodpecker-ci.org/repos/3780/pipeline/33744/38
• https://ci.woodpecker-ci.org/repos/3780/pipeline/33738/38

flaky-ness is gone:

[woodpecker-bot]

Surge PR preview deployment was removed

[6543]

sorry i tested this as hard as i could .. by running manual real pipelines in local and docker backend, rerun tests suite a lot and catch all imposible bugs etc ... still I should have just tested "services" in a real pipeline. i did not :/

-> #6507

this here make sure traces are uploaded before we return, but we depend on return to fully stop the workflow ATM ...

the mentioned pull makes workflow shutdown explizite and still has defer for cases we have an error etc...