fix: e2e - docker logs fix (#2032)

There was an issue where containers were removed before the docker logs
step. Containers are now kept around until GH actions are done

fixes #2031

tests/docker-scripts/run.sh

@@ -4,10 +4,19 @@
 # The argument should contain the container name to run as the first argument (e.g. web-ui-tests)
 # Once the container finishes its task, this script shuts down the docker containers
 # because docker compose run starts dependent containers, but does not stop them
-
 pushd "$(dirname "$0")" # Set pwd to this directory
-docker compose run --service-ports --build --rm "$@" # Passes all arguments to the compose file
-exit_code=$?
-docker compose down
+
+if [[ "${CI}" == "1" || "${CI}" == "true" ]]; then
+ # In CI, keep the container in case we need to dump logs in another
+ # step of the GH action. It should be cleaned up automatically by the CI runner.
+ docker compose run --service-ports --build -e CI=true "$@"
+ exit_code=$?
+ docker compose stop
+else
+ docker compose run --service-ports --rm --build "$@"
+ exit_code=$?
+ docker compose down
+fi
+
 popd # Reset pwd
 exit $exit_code