workflow troubleshooting

[vd1992]

trying to troubleshooting the docker/e2e issues...

okay, each commit represents a solution to the recent issue of workflows for smoketest and cypress
both appear to resolve the issue, I am partial to the latter since I do not know if there is a purpose to the postrun save, disabling it solves the issue for sure and shaves some time off test runs

related to docker layer caching action: satackey/action-docker-layer-caching#139 (comment)

per conversation here
https://talent-cloud.slack.com/archives/C0259G6KXJ6/p1657839973910039

[vd1992]

looking at some previous runs, this one for example, post-run took 6 minutes
https://github.com/GCTC-NTGC/gc-digital-talent/runs/7346142902?check_suite_focus=true

[petertgiles]

What are the "docker/e2e issues"? I'm not sure what this branch is solving. Thanks for the link.

[patcon]

Nice! If I'm understanding correctly, 8782930 looks like it also resolved it by clearing up space on the VM? This feels like the least change to get things working.

(I believe that the skip-save: true alternative totally turns off updating of the cache, which is the same as removing the cache step.)

Preference is for the first solution 8782930, since it preserves caching. But if you do go with the skip-save option (because I'm missing some consideration), I'd propose adding a comment and maybe a TODO to "re-enable once PR #### is resolved", as my experience is that temporary work-arounds get forgotten if not annotated in code :)

[patcon]

Preference for solution in prior commit, unless I'm missing some consideration :)

[vd1992]

Preference for solution in prior commit, unless I'm missing some consideration :)

aye aye

[patcon]

Thanks!

[vd1992]

Thanks!

workflows please cooperate now!