Retry docker image pull on failure

[palfrey]

Periodically we get a failed build because ghcr.io was a bit erratic. I suspect a retry of the pull with a few seconds delay would have worked.
e.g.

Unable to find image 'ghcr.io/renovatebot/renovate:39' locally
39: Pulling from renovatebot/renovate
docker: Get "https://ghcr.io/v2/renovatebot/renovate/manifests/sha256:c186f5db05148fa5fd0bba339c09012762c31138163ebac947e782546d5158b0": dial tcp 140.82.114.33:443: i/o timeout.
See 'docker run --help'.
Error: The process '/usr/bin/docker' failed with exit code 125

[rarkins]

Is it always the same message and error code?

[palfrey]

I thought I'd seen more than one, but apparently not! Or at least I now can't find any logs with this :(

[viceice]

it looks like an issue in your infrastructure, you could pull the image before the action in your workflow as workaround.