pages:deploy job restarts and fails without intervention

Summary

I have a page job that creates the automatic pages:deploy job. The latter takes a long time and then is retried on its own and fails, while the initial instance apparently keeps running:

All previous stages went fine, note the bottom job is marked as "running" (I've refreshed the page several times), the one directly above was retried without my intervention, nothing happend to trigger it, and it failed with "missing pages artifacts" in the pipeline view. The pages artifact (public folder) was generated and uploaded successfully by the pages job.

Steps to reproduce

No idea. I'm trying to deploy a doxygen build that is quite large, but should be below 1 GB, and I'm using shared runners.

Example Project

https://gitlab.com/Jellby/lapack-release/pipelines/9604838/builds

https://gitlab.com/Jellby/lapack-release/pipelines/9605981/builds

What is the current bug behavior?

pages:deploy retries by itself and fails. Or the initial job is stuck as "running" forever, since I can't see what's going on there.

What is the expected correct behavior?

It should wait until it's finished, with success or failure, and with a proper message or log.

Output of checks

This bug happens on GitLab.com