#github-actions (2022-07)

is the goal to save minutes or looking to have the workflow “pause/resume”?

to save minutes

GHA workflow invokes Google Cloud Function -> GHA workflow keeps chugging along/ends

Google Cloud Function runs Firebase bits -> Triggers GHA workflow to do the rest of whatever was happen

it’s not unified, but it gets the job done

I think the feature request to create a on-demand callback is cool, but I also don’t see a clean way for GH to implement that where the compute you’re using isn’t still burning minutes

though maybe you could finagle the manual approval step

trigger cloud function w/workflow run ID as an arg and it runs the firebase bits then does a GH API call to approve the workflow to keep running?

Reviving zombies here, but why not POST an authenticated workflow_dispatch?

Alternatively, if costs is the ultimate concern, you might compare the cost of a small self hosted runner

Thanks for the suggestion @Jim Park

I guess that could work, one concern is the detachment between jobs and getting easy visibility of whether the “entire pipeline” completed.