Skip to main content

Kubernetes Reference

The end-to-end install sequence lives in the Production Deployment Runbook. This page is the kubectl-level companion: the workloads the platform runs, and the commands to inspect and diagnose them once deployed. The runbook is authoritative for install and configuration; use this page to check a running deployment.

The platform requires a CNCF-conformant Kubernetes distribution. We support Amazon EKS (k3s for local development and evaluation only). Other distributions such as GKE or AKS are not validated and are not supported by default; contact Connamara to discuss one (see Version Compatibility). Where a cloud-specific choice arises (managed database, ingress controller, container registry), substitute the equivalent service from your provider.

What the platform runs

After a successful install, the namespace holds these workloads:

WorkloadKindExpected statusDefault count
conncentric-orchestrator-*DeploymentRunning2
conncentric-adapter-*DeploymentRunning4
conncentric-portal-*DeploymentRunning1

The install also runs two short-lived Jobs, a schema migration before the application pods and an installer afterwards. Both remove themselves on success, so a namespace with no Jobs left is the expected end state; either one is retained when it fails, which is when its log is worth reading.

Adapter pods that hold no lease report 0/1 Ready. That is expected: an idle adapter is a standby, not a fault. See Deployment Topologies.

Inspect a running deployment

The deploy/… and job/… names below track your Helm release name; if yours is not conncentric, run kubectl get deploy -n conncentric first to find the actual names.

# Workloads and pod health
kubectl get pods -n conncentric

# Logs for a specific pod (tail)
kubectl logs <pod-name> -n conncentric --tail=200

# Installer job result, available only if it failed (it removes itself on success).
# Pass 1 applies the base Essentials distribution and enabled plugins; Pass 2 applies any custom bundles from installer.customBundleUrls.
kubectl logs -l app.kubernetes.io/component=installer -n conncentric

# Orchestrator health (expect "status":"UP")
kubectl exec -it deploy/conncentric-orchestrator -n conncentric -- \
curl -sf http://localhost:8080/actuator/health

# Ingress routing (Portal reachability)
kubectl describe ingress -n conncentric

Troubleshooting

SymptomLikely causeAction
Orchestrator pod in CrashLoopBackOffDatabase unreachable or wrong credentialsCheck database.* values; test DNS from a pod: kubectl exec -it <pod> -n conncentric -- nslookup <db-host>
Portal loads but shows "Unable to reach the server"Ingress not routing /api to the Orchestratorkubectl describe ingress -n conncentric and confirm the backend and path
Adapter pods running but no connectors or components offered in the PortalInstaller job did not completeA failed installer job is retained, so read its log with kubectl logs -l app.kubernetes.io/component=installer -n conncentric; re-run with helm upgrade conncentric oci://<your-registry>/conncentric/charts/conncentric --version <release-tag> -n conncentric -f values.production.yaml

For OIDC sign-in failures see Authentication; for broader diagnostic patterns see Troubleshooting.

See also