Deployment healthcheck message skips unhealthy services that aren't reporting metrics
During auto-deploys we allow a packaage to 'bake' on the gprd-cny environment for a period of time. Once the baking time completes we run a production health check to determine if there are blockers that should prevent a promotion to Production. Services with unhealthy metrics cause the check to fail and block deployments. But, we're missing the failure case of a service not reporting metrics due to unhealthiness, if this happens the the production health check skips over the service rather than reporting it as unhealthy:
versus the check running again 2 minutes later
If a service is missing metrics We should assume it is unhealthy and report it as such to prevent deployments