The easiest workaround is to rename the container, assuming you're running some docker build+run or docker-compose.
Depending on your workflow, consider post-test removal/pruning in particular with cancelled(), failure() or always(). If you're using self-hosted runners, login to that system and kill the container. For Github-hosted runners, I don't think there is any access to them. You'll have to wait until the default timeout kills the container.