Add configuration for optional CPU request on build pod

[bluek1te]

Add ability to configure CPU request on build pod - We have cluster restrictions that require us to affix a cpu request before spawning pods so this would help us immensely.

[manics]

Unfortunately this won't work, because the "build pod" doesn't build the image directly, it calls out to an external Docker or Podman build daemon.

[bluek1te]

@manics
We just need a CPU resource attribute to be affixed to the build pod. Even if it calls out to an external docker daemon, the "build pod" still takes some amount of CPU to run. There are already configurations for memory requests/limits for this specific pod, so I don't see why this would be any different.

[bluek1te]

Bump

[rgaiacs]

@bluek1te some tests are failing and I think we might have fixed it already. Can you rebase this pull request so that we can check the tests? Thanks!

[bluek1te]

@rgaiacs looks like the tests failed after syncing, but they are timeout errors and I don't think they are related to my changes. Is there an easy way to rerun these tests on my side without pushing empty commits?

[rgaiacs]

Is there an easy way to rerun these tests on my side without pushing empty commits?

I recently discovered that close and open the pull request will trigger a new CI job.

There is also an option when you look at the logs of the job but I don't know what access level is required to trigger the re-run.

I'm manually trigger a re-run of the failed jobs.

[rgaiacs]

During the first CI run, tests (v1.30, main) and tests (v1.29, helm, pink failed.

During the second CI run, only tests (v1.30, main) failed.

The reason for the tests to fail are different. The first time was

=========================== short test summary info ============================
FAILED binderhub/tests/test_repoproviders.py::test_gitlab_ref[v10.0.6-b3344b7f17c335a817c5d7608c5e47fd7cabc023] - tornado.curl_httpclient.CurlError: HTTP 599: Operation timed out after 20000 milliseconds with 0 bytes received
FAILED binderhub/tests/test_repoproviders.py::test_gitlab_ref[HEAD-True] - tornado.curl_httpclient.CurlError: HTTP 599: Operation timed out after 20000 milliseconds with 0 bytes received
FAILED binderhub/tests/test_repoproviders.py::test_gitlab_ref[nosuchref-None] - tornado.curl_httpclient.CurlError: HTTP 599: Operation timed out after 20001 milliseconds with 0 bytes received
===== 3 failed, 147 passed, 3 deselected, 12 warnings in 183.96s (0:03:03) =====

the second time was

=========================== short test summary info ============================
FAILED binderhub/tests/test_repoproviders.py::test_figshare[10.6084/m9.figshare.9782777.v1-10.6084/m9.figshare.9782777.v1-9782777.v1-https://doi.org/10.6084/m9.figshare.9782777.v1-figshare-9782777.v1] - tornado.httpclient.HTTPClientError: HTTP 403: Forbidden
FAILED binderhub/tests/test_repoproviders.py::test_figshare[10.6084/m9.figshare.9782777-10.6084/m9.figshare.9782777.v1-9782777.v1-https://doi.org/10.6084/m9.figshare.9782777.v1-figshare-9782777.v1] - tornado.httpclient.HTTPClientError: HTTP 403: Forbidden
===== 2 failed, 148 passed, 3 deselected, 12 warnings in 131.66s (0:02:11) =====

I don't think this pull request is related with the above test failure given that only the test for figshare failed.

[bluek1te]

I've rerun the tests and they look like they are all passing now.

[bluek1te]

Bump

[bluek1te]

@manics

[rgaiacs]

Looks good to me.

[rgaiacs]

@bluek1te Thanks for the contribution. Have you heard about our roadmap workshop? Maybe you want to attend and contribute?

Let's wait a few more days to get a second review on this pull request.

I added a reminder to myself to get this done by next week.

[minrk]

Sorry for the wait. This has been addressed with a more general resources in #2071 which allows you to set arbitrary resource requests/limits on the repo2docker pod.