[maintenance] Adjust test job timeouts in GitHub Actions CI/CD jobs

[webknjaz]

🐞 Describe the bug

The CI doesn't complete currently because the test jobs are capped to run under 15 minutes and they now get killed on timeout. I believe this is called by the new pip resolver doing backtracking which is rather time-consuming.

💡 To Reproduce

It's visible in any recent CI build.

💡 Expected behavior

The jobs should complete. Carefully bumping the timeout in the workflow file should be enough.

📋 Logs/tracebacks

Example: https://github.com/aio-libs/aiohttp/runs/1669008218?check_suite_focus=true

📋 Your version of the Python

N/A

📋 Your version of the aiohttp/yarl/multidict distributions

repo master

📋 Additional context

N/A

[greshilov]

Increased timeout will not solve the problem completely, I guess. Stricter requirements.txt could help though.

[webknjaz]

I think increasing the timeout should be enough to get it to run completely for the first time and then many things will probably get picked up from cache. One of the problems there is that when pip has to download many package versions, it makes the backtracking substantially slower when those packages don't have wheels.