With CNG pipeline being a faster alternative to the Omnibus build, and with the Omnibus build not being run in MRs by default. It could be beneficial to run the `ai-gateway`-related tests in the CNG pipeline. Currently the `ai-gateway` job runs tests tagged with `:ai_gateway` against the `latest` ai-gateway in mock response mode. ai-gateway has a helm chart: https://gitlab.com/gitlab-org/charts/ai-gateway-helm-chart Note: i notice at the above link `⚠️ This helm chart is meant for self-hosted models users only!`. Although we would be using it in with `AIGW_MOCK_MODEL_RESPONSES=true` so I am not sure if it really matters what model we are using.