https://github.blog/2022-05-09-supercharging-github-actions-with-job-summaries/ may be helpful for the actions collected in the CD workflow, rather than just echoing a cryptic message to the log. Currently it is very hard to figure out from

what is going on; since we cannot limit check_run triggers we get dozens of promptly aborted runs and the single one which you expected to deploy but which did not is hard to find in the noise. Clicking on one run does not help much either:

until you dig down into the logs: