503 error received but private IP works for backend target

[RcardonaFinch]

Hello,

I am receiving a 503 error when I use a hostname for my backend target. But when I use the private IP address, it works just fine. I am running nginx (for testing) on a dedicated GKE cluster (headless service) which is successfully integrated with Cloud DNS. I've verified that any pod changes on GKE are being automatically updated to Cloud DNS. In addition, I can successfully resolve the hostname locally in the cluster. 

I performed a trace in Apigee and found that the hostname is not resolving for the backend target. What are the steps involved with verifying that Apigee is successfully configured to use Cloud DNS for resolution? 

503 error when using hostname (pulled from Cloud DNS) for backend target:

Postflow Target:

Cloud DNS A record:


Hostname resolved successfully in GKE Cluster

Your help is appreciated. 🙂

[markjkelly · Accepted solution]

Hi @RcardonaFinch - Have you configured DNS Peering? It sounds like that may be the cause of the issue.

Your private DNS zones are not resolved by Apigee runtime by default. With DNS peering, you can allow the Apigee runtime resolve those private DNS records. See here for more info and steps on how to do this.

[RcardonaFinch]

Hi @markjkelly, thank you for that.

When running the peering command, I am receiving an error message which I am having trouble interpreting. Could you please assist?

[RcardonaFinch]

@markjkelly Please disregard. This happened because my DNS suffix was missing a period in the end so it did not locate that suffix. The DNS peer was successfully created.

[RcardonaFinch]

I've now tested this successfully. Thanks for the help @markjkelly