GitHub Actions error: `Error: fetch failed`

[gaggle]

Got this all over my CI:

It ran successfully when I retried the job. Is there retry logic to consider adding to increase fetching resilience?

(and thanks for making this amazing project available)

[jhheider]

This seems like a good idea. GitHub actions seems to suffer intermittent failures of connectivity even during runs. pantry's closed issues are a testament to this.

[mxcl]

seems like a good plan yeah. three reties perhaps with a small sleep?

[jhheider]

sleep 1; sleep 3; sleep 5? probably not worth including any kind of sophisticated expo backoff tech.

[mxcl]

here we had a fetch hang for 2 hours:

[jhheider]

Yep. Sometimes an action step will hang for 6 hours then time out. I get github status texts most days of the week that some aspect of their infrastructure is degraded. :/

[mxcl]

oh so not likely because our code sucks?

[jhheider]

Anything is possible, of course.

[mxcl]

v3 is a rewrite that shouldn’t have these issues. Provided rust is more reliable in pkgx itself that is.