@kapa.ai I have deployed a Self-Managed Community version Airbyte Platform app v1.7.1 in my GKE cluster. I created a postgreSQL connector earlier then deleted with the option "no" for deletion of connection. However, connection was deleted. But I am not able to recreate it and getting following error "An unexpected error occurred. Please report this if the issue persists. (HTTP 500)".

There are a couple of known issues with Airbyte v1.7.1 on Kubernetes that could be contributing to what you're experiencing. Likely Causes 1. Ingress Misconfiguration (Most Probable) Version 1.7.1 introduced a breaking change in how traffic needs to be routed. If your ingress is not correctly configured, API calls fail with 500/404 errors. Specifically, traffic to

/api/v1/connector_builder

must be routed to the

airbyte-airbyte-connector-builder-server-svc

service, while other traffic goes to the webapp service. [Ingress 1.7 docs] For NGINX Ingress, the correct configuration looks like this:

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: airbyte-ingress
spec:
  rules:
    - host: airbyte.your-domain.com
      http:
        paths:
          # Routes builder API traffic to the correct service
          - path: /api/v1/connector_builder
            pathType: Prefix
            backend:
              service:
                name: airbyte-airbyte-connector-builder-server-svc
                port:
                  number: 80
          # Fallback for all other UI traffic
          - path: /
            pathType: Prefix
            backend:
              service:
                name: airbyte-airbyte-webapp-svc
                port:
                  number: 80

[Ingress fix example] 2. Workload API Authentication Issue There is also a known platform-level issue where connector-sidecar containers don't receive the necessary authentication tokens to communicate with the workload API server, resulting in HTTP 500 errors like

DocStoreAccessException: Unable to write output

. This has been reported by other users on v2.x but may affect your setup as well. [GitHub Issue] Recommended Steps 1. Update your ingress rules per the v1.7 ingress documentation — this is a required change for v1.7.1. 2. Upgrade to a newer patch if possible. Version 1.7.6 fixed an

Internal Server Error: not yet implemented

error that caused various navigation and connector issues. [v1.7 patches] 3. Check browser network tab for the exact failing API endpoint and HTTP status to narrow down the root cause. 4. Restart the

airbyte-server

pod — in at least one similar case, a server restart resolved a 500 error caused by a cache problem. [GitHub Issue] The ingress misconfiguration is the most commonly reported root cause for HTTP 500 errors when recreating connectors on v1.7.1 Kubernetes deployments. 👍 0 button 👎 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.