SYMPTOM

The user encountered an issue where the deployment of a new Flex Gateway on a Kubernetes cluster failed with a 503 Service Unavailable error. The error occurred during the execution of the Helm upgrade command, and the gateway was not starting as expected. The Runtime Manager Flex Gateway UI shows running and connected but describing the pod shows that the pod is not ready

$> helm -n gateway upgrade -i --create-namespace --wait ingress flex-gateway/flex-gateway --set-file registration.content=registration.yaml --set gateway.mode=connected

Warning Unhealthy 2s (x15 over 2m2s) kubelet Readiness probe failed: Error: probe check failed: {"name":"Service Unavailable","message":"Gateway not started","status":503}

CAUSES

The helm upgrade --wait flag will wait for a working API configuration

​​​​​​​SOLUTION

1. deploy an API for the Gateway to become ready OR
2. run the helm command without the --wait

https://docs.mulesoft.com/gateway/latest/flex-gateway-k8-getting-started#deploy-flex
https://docs.mulesoft.com/gateway/latest/flex-gateway-k8-getting-started#troubleshooting-k8