I have a gitlab job in which I am running some cypress tests against a jboss and a postgresql. Both jboss and postgresql are declared as services in the job. Here is the job’s relevant part:

e2e-tests:
  stage: test-images
  before_script: []
  services:
  - name: ourdockerrepo/bitnami/postgresql:16-debian-12
    alias: db
  - name: ourdockerrepo/jboss-app-to-test:something
    alias: eap
image:
  name: cypress/included:13.6.4
  entrypoint: [ "" ]
script:
  - |
    start=$EPOCHSECONDS
    max_seconds_timeout=600
    while [ $(wget --spider -S "http://eap/app" 2>&1 | grep "HTTP/" | grep -q 200 && echo "true" || echo "false") == "false" ];
    do
      echo ".";
      sleep 1;
      if (( EPOCHSECONDS-start > max_seconds_timeout )); then echo "timeout! eap not raised within $max_seconds_timeout seconds."; exit 1; fi
    done
  - npm install
  - npm run (cypress stuff)
...

I am experiencing that gitlab’s healthcheck for both services expire and they both go in some sort of race condition. Sometimes db starts before eap and tests run as expected. Sometimes eap fails to start properly because it came up before db does, causing script’s while check to timeout. I am a gitlab greenhorn, is there any way to prevent this behavior and start services in the order I declared in job? Thankyou!