add k8s log collector to CI

[willcl-ark]

Currently, it's very painful to debug inconsistencies in the CI vs a local environment. Additionally, when things go wrong, we don't get any kubernetes output at all.

Fix this by adding a job which uploads artefacts to GH for each job.

stern allows tailing of all pods (in all namespaces), and can dump to json (I chose default mode here).

See an example of the artefacts here: https://github.com/willcl-ark/warnet/actions/runs/10763746456

[josibake]

ACK, +1 for making CI failures easier to debug. I don't have any expertise in the GA area, but this seems reasonable to me.

(I chose default mode here

why not founder mode?

[willcl-ark]

ACK, +1 for making CI failures easier to debug. I don't have any expertise in the GA area, but this seems reasonable to me.

(I chose default mode here

why not founder mode?

Ok, I'm going full founder mode. Just ordered one of these in each colour for each of us: https://levelsio.com/products/founder-mode-hat-100-cotton

[bdp-DrahtBot]

The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.

Conflicts

No conflicts as of last run.

[josibake]

needs a rebase and would be great to get someone else's opinion who is more familiar with GA / our CI, cc @pinheadmz

[josibake]

LGTM and seems useful, if this ends up being problematic for GA we can always revert.