ci(build-and-test-differential): use github runner

[kminoda]

Description

Use GitHub runner.

The runner has been changed as follows:

• chore(workflows): update build-and-test-differential to use self hosted runners #4364 GitHub runner -> self-hosted runner
• chore(ci): use ec2 for build-and-test-differential #4413 self-hosted runner -> EC2 runner
• chore(ci): revert "use ec2 for build-and-test-differential (#4413)" #4477 EC2 runner -> self-hosted runner

Apparently the disk image size has been reduced, and thus we would like to use GitHub Runner again, since current self-hosted runner based CI is too slow to execute multiple PRs.

Test run: https://github.com/HansRobo/autoware.universe/actions/runs/7736637727/job/21094318337
Remaining disk space:

$ dh -a

Filesystem     1K-blocks     Used Available Use% Mounted on
overlay         87204404 75316916  11871104  87% /

du results in the latest full-built Autoware:

$ cd /path/to/pre-built/autoware
$ du -d0
11311632        .  

Tests performed

As shown above

Effects on system behavior

There is a risk in a near future that the CI will fail again due to the disk usage limit as it was before.

Autoware developers need to understand the importance of reducing the file size of both scripts and the generated binaries.

Pre-review checklist for the PR author

The PR author must check the checkboxes below when creating the PR.

• I've confirmed the contribution guidelines.
• The PR follows the pull request guidelines.

In-review checklist for the PR reviewers

The PR reviewers must check the checkboxes below before approval.

• The PR follows the pull request guidelines.

Post-review checklist for the PR author

The PR author must check the checkboxes below before merging.

• There are no open discussions or they are tracked via tickets.

After all checkboxes are checked, anyone who has write access can merge the PR.

[shmpwk]

LGTM because Github Actions perfomance was improved.
https://github.blog/2024-01-17-github-hosted-runners-double-the-power-for-open-source/

[xmfcx]

$ dh -a

Filesystem     1K-blocks     Used Available Use% Mounted on
overlay         87204404 75316916  11871104  87% /```

@kminoda it's better to run df -h since it gives disk info in a more readable way.

@mitsudome-r this was never supposed to be merged because even in the PR description, the available space is 11871104 KB which is 11.87 GB.

The available space for public runners was always this size.

This PR was merged, worked a couple of times randomly. Because by nature, the build and test differential workflow can build varying count of packages, changing the size footprint.

And the container size was another blocker, sometimes preventing the job from starting in the first place.

As for the updated workflow:

• ci: update build-and-test and -differential jobs #6471

from the latest job:

Run df -h
Filesystem      Size  Used Avail Use% Mounted on
overlay          73G   61G   12G  85% /

The available space for the containers is still the same.

[xmfcx]

I've also browsed some passing jobs that has passed from this PR's merge.

Without CUDA

https://github.com/autowarefoundation/autoware.universe/actions/runs/7757485516/job/21157041781#step:4:6

Filesystem      Size  Used Avail Use% Mounted on
overlay          84G   64G   20G  77% /

With CUDA

https://github.com/autowarefoundation/autoware.universe/actions/runs/7757485516/job/21157042168#step:4:6

Filesystem      Size  Used Avail Use% Mounted on
overlay          84G   74G   11G  88% /

It seems you got some 84G size machines where it worked. But when it failed, it was with 73G machines and CUDA version of the docker image didn't fit in there.