Thank you for the report!

Those chroots which already started really flip to the failed state now. They are simply interrupted by a background "kill" process. This is expected, we changed the behavior
to make the make the databas more clear and this was simple to implement.

Does this make any real troubles, or is that just cosmetics? The overall build state
should be canceled I believe.

Its really just cosmetic.
It should be noted that it seems to be rather inconsistent, when a build with a failed chroot is canceled the status of the chroot becomes canceled instead of staying failed

Ok, let's call this a feature. I admit it would be nicer to have the canceled state
also in the chroots with "killed" build logs. It is a matter of teaching
copr-rpmbuild to return special exit status that denominates "canceled", and
propagating that exit status back to frontend.

The change I talked about is 65b3435 and 991831f, and that
were really needed and good changes. What we see here is a cosmetic leftover
which is not 5-minute work to fix.

This issue has been migrated to GitHub: https://github.com/fedora-copr/copr/issues/1622