Recreating CI jobs from upstream used for release signal

[dims]

We have patches against upstream components including Kubernetes, It would be good to adopt some of the CI signals used by k8s release teams as well. I see we do run conformance tests, but those are do not cover the range of functionality and possibilities of things that could go wrong.

Upstream release team uses the following jobs for most of the time:

• https://testgrid.k8s.io/sig-release-master-blocking
• https://testgrid.k8s.io/sig-release-master-informing

Around the time when a branch is cut, they copy over the jobs like this:

• https://testgrid.k8s.io/sig-release-1.26-blocking
• https://testgrid.k8s.io/sig-release-1.26-informing

There some some kubelet related tests that are also important (they run the tests marked with [NodeConformance]:

• https://testgrid.k8s.io/sig-node-release-blocking

(this exercises the containerd/runc/kubelet trio)

Not sure if we do these any of these, but filing to start the conversation! We don't have to do all of these either, we can pick and choose / prioritize, may be do the easy ones first? :)

thanks,
Dims

[danbudris]

Love this, thanks for filing! We've been working on improving our test framework and looking to expand our coverage, so this fits nicely into that work.

[markapruett]

First step of this would be to have an engineer look into which tests we should pull in and how much lift it would take.