rootfs:overlays:fluster Add verbose option for fluster parser

[denisyuji]

This is a continuation of the PR #2609 accidentally closed by me.

This change adds --verbose option for fluster parser to make fluster more verbose to facilitate debug. When --verbose is provided, the fluster_parser.py script will forward it to fluster which will enable the verbosity of GStreamer. The GStreamer verbosity facilitates the debug on regressions investigations.
The low verbosity level enabled by fluster adds few hundreds of lines on logs when this option is provided, without significant impact on the performance. Anyways, the verbosity option should be disabled by default on fluster_parser.py and will not be used by kernel-ci pipeline by default.
The intention is to enable users to add --verbose flag manually only when investigating a codec regression.

[laura-nao]

LGTM! @nuclearcat given the changes are fairly trivial, should we build a new bookworm-gst-fluster image directly for production without going through staging?

[pawiecz]

should we build a new bookworm-gst-fluster image directly for production without going through staging?

Agreed, these changes could prove themselves useful in shipping fluster features quicker (no need to change rootfs URLs in templates/configs).

From scheduler entries in kernelci/kernelci-pipeline#705 I see that the main target are arm64 platforms - I triggered a build pipeline for them.

[pawiecz]

Image rebuild consistently fails, I posted details on the thread in a Slack channel

[pawiecz]

Root cause of the image build failure has been found and patched - staging build artifact.

I'll rebuild the image for production when the patches above get merged.

[pawiecz]

Patches were merged, building bookworm-gst-fluster images is back to normal since 20240819.0.

@denisyuji please rebase this PR if you need -v available in fluster_parser.py for debugging/investigations - I'll trigger image creation.

[denisyuji]

Thank you @pawiecz , rebase is done

[pawiecz]

Updated images are available but I wasn't able to verify them by reusing TestJob definitions from related PRs (details)