The podman test failure remains after #485 that solved some other failures, but is tolerated with continue-on-error: true on the step running podman tests.
The failure is that the podman system service that starts containers, and is started by the non-root user, doesn't have sufficient cgroup permissions to constrain CPU usage etc in the started containers I think. This is described in podman's troubleshooting docs.
docker.errors.APIError: 500 Server Error for http+docker://localhost/v1.40/containers/65b02b7e1a0aa53a877e961e1f65b3af4be0c89c20fe4d9bc4d2bc619f816002/start: Internal Server Error ("the requested cgroup controller `cpu` is not available: OCI runtime error")