docker: Error response from daemon: Conflict

Docker container name conflict in CI is almost always caused by parallel pipeline execution. GitHub Actions matrix strategies, Bitbucket parallel steps, and Jenkins parallel stages can run multiple jobs simultaneously on the same Docker host. If each job launches a container with the same name (hardcoded in docker run --name my-app or defined as a fixed container_name in docker-compose.yml), the second job fails with "container name already in use". The fix for docker run: omit --name entirely — Docker generates a unique random name like dreamy_liskov. For docker-compose.yml, remove explicit container_name properties, which override Docker Compose's default naming (that includes the project name and service name, making them unique per directory). For pipelines that require a fixed name, add a docker rm -f container-name || true step before the docker run step. The || true prevents the cleanup step from failing when the container does not exist on the first run. Docker Compose handles parallel runs better: by default it names containers as <project>_<service>_<index> where <project> defaults to the directory name. Running docker-compose -p unique-$BUILD_ID up with a pipeline-specific project name (using the commit SHA or build number) namespaces all containers in a pipeline run uniquely, completely eliminating cross-job conflicts.