Troubleshooting the Installation

Start with the External DNS logs, where most problems show up:

kubectl -n tetrate-gslb-system logs deploy/tetrate-gslb-external-dns-local -c external-dns

Symptom	Cause and fix
InvalidCharacterString (Value should be enclosed in quotation marks)	dnsProvider isn't set to aws. Every DNS write is being rejected.
Records appear in the wrong region's zone	zoneIdFilters missing or wrong. Set it on both instances in every cluster.
AccessDenied on ChangeResourceRecordSets	The IAM policy doesn't cover that zone, or IRSA isn't active — check AWS_ROLE_ARN on the pod as in Step 3.
Throttling: Rate exceeded	Route53 allows ~5 writes/second per account across all clusters. Raise reconcileRequeueSeconds and the two *TXTRecordHealthyDurationSeconds values.
Lookups return NXDOMAIN inside a cluster	The cluster's VPC isn't associated with the hosted zone, or the VPC has DNS support disabled. Recheck Step 1.
A gateway never appears in DNS	Check its Service has a load balancer hostname, that the hostname resolves from inside the cluster, and that the Service has ready endpoints.
Pods crash with exec format error	The images are arm64 but the nodes are amd64. Refer to the Acquire Images process.

The general troubleshooting steps in Troubleshooting apply too.