Skip to main content

Authentik on GKE Autopilot — Lab Guide

📖 Configuration Guide

Overview​

Estimated time: 60–90 minutes

authentik is an open-source identity provider — single sign-on via OIDC and SAML, LDAP, SCIM, MFA, and proxy authentication; a self-hosted alternative to Okta, Auth0, and Keycloak. This lab takes you through the full operational lifecycle of the Authentik on GKE Autopilot module on Google Cloud: deploy it, access and verify it, run it day-to-day, observe it, diagnose common problems, and tear it down.

The lab focuses on operating the GKE module and the Google Cloud platform, not on authentik product features. For the complete list of provisioned services and every configuration input (organised by group), see the Configuration Guide — this lab deliberately does not duplicate that detail so it stays accurate over time.

Objectives​

By the end of this lab you will be able to:

  • Deploy the module from the RAD platform and locate the resources it provisions.
  • Connect to the GKE cluster, verify the workload, and log in as the bootstrapped akadmin.
  • Create a first OIDC application and provider in the authentik UI.
  • Perform day-2 operations — inspect pods, update the version, watch the co-located worker, and open a database session.
  • Observe the workload with Cloud Logging and Cloud Monitoring.
  • Diagnose and resolve the most common deployment and runtime issues.
  • Tear the deployment down cleanly.

Prerequisites​

  • Services_GCP deployed in the target project (provides the VPC, GKE Autopilot cluster, Cloud SQL, Artifact Registry, and shared service accounts this module depends on).
  • A Google Cloud project with billing enabled.
  • gcloud CLI and kubectl installed; gcloud auth login and gcloud auth application-default login completed.
  • Project Owner (or equivalent) IAM on the project.
  • RAD platform access with permission to deploy modules into the project.

Set these shell variables once; every task below reuses them:

export PROJECT="<your-gcp-project-id>"
export REGION="us-central1"          # the region you deploy into

Task 1 — Deploy the module [Automated]​

  1. In the RAD platform, open Authentik (GKE), set project_id, and review the inputs. Configure only what you need — the Configuration Guide documents every input by group, with defaults. Review the estimated cost (if credits are enabled) and click Deploy, which opens the deployment status page with real-time logs.

  2. The platform provisions the Kubernetes namespace and Deployment, a Cloud SQL (PostgreSQL 15) database with its Secret Manager secrets (AUTHENTIK_SECRET_KEY, the akadmin bootstrap password, and the database password), a Cloud Storage media bucket mounted at /media via GCS Fuse, builds the thin custom image (FROM ghcr.io/goauthentik/server), and runs a one-shot database-initialisation Job. No Redis is provisioned — authentik ≥ 2025.10 keeps cache, sessions, and its task queue in PostgreSQL. First deploys take roughly 20–35 minutes (Cloud SQL creation dominates); authentik then runs its full migration suite on first boot, so allow a few extra minutes before the pod turns Ready.

  3. When it completes, connect to the cluster and discover the resources with name-agnostic filters:

    CLUSTER=$(gcloud container clusters list --project="$PROJECT" \
      --format="value(name)" --limit=1)
    gcloud container clusters get-credentials "$CLUSTER" \
      --region="$REGION" --project="$PROJECT"

    NAMESPACE=$(kubectl get ns -o name | grep authentik | cut -d/ -f2)
    kubectl get pods,svc -n "$NAMESPACE"

    SERVICE_IP=$(kubectl get svc -n "$NAMESPACE" \
      -o jsonpath='{.items[?(@.spec.type=="LoadBalancer")].status.loadBalancer.ingress[0].ip}')
    SERVICE_URL="https://${SERVICE_IP}.nip.io"
    echo "Namespace: $NAMESPACE"
    echo "URL:       $SERVICE_URL"

    (The deployment outputs also report namespace, service_external_ip, and service_url directly.)

Task 2 — Access & verify [Manual]​

  1. Confirm the workload is healthy. authentik exposes an unauthenticated readiness endpoint that returns 200 only once migrations have completed and the database is reachable:

    curl -sk "$SERVICE_URL/-/health/ready/" -o /dev/null -w '%{http_code}\n'   # expect 200
    curl -sk "$SERVICE_URL/-/health/live/"  -o /dev/null -w '%{http_code}\n'   # expect 200

    A 200 status alone is not proof the server answered — also confirm the response body is non-empty (a 200 with a zero-byte body means the wrong process answered :9000; see Task 5):

    curl -sk "$SERVICE_URL/" | wc -c   # expect non-zero — the login page HTML

  2. Retrieve the bootstrapped admin credential from Secret Manager. The module creates the built-in akadmin user on first boot with the configured bootstrap_email (default admin@techequity.cloud) and this password:

    BP_SECRET=$(gcloud secrets list --project="$PROJECT" \
      --filter="name~authentik AND name~bootstrap-password" \
      --format="value(name.basename())" --limit=1)
    gcloud secrets versions access latest --secret="$BP_SECRET" --project="$PROJECT"; echo

  3. Open $SERVICE_URL in a browser and sign in with username akadmin and the password from step 2. (If the bootstrap variables were absent on the very first boot, authentik instead offers the initial-setup flow at $SERVICE_URL/if/flow/initial-setup/ — set the admin password there.)

  4. Create a first OIDC application and provider in the UI:

    1. Open the Admin interface (top-right avatar → Admin interface, or $SERVICE_URL/if/admin/).
    2. Go to Applications → Applications and click Create with Provider (the wizard creates the application and its provider together).
    3. Name it (e.g. demo-app), choose OAuth2/OIDC Provider, accept the default authorization flow, and set the client's Redirect URI to your test application's callback URL.
    4. Finish the wizard, then open the provider to copy the Client ID, Client Secret, and the OIDC endpoints (also discoverable at $SERVICE_URL/application/o/<slug>/.well-known/openid-configuration).

Task 3 — Operate & keep it running (Day-2) [Manual]​

  1. Inspect the workload:

    kubectl get deploy,pods,svc,hpa -n "$NAMESPACE"
    kubectl describe deploy -n "$NAMESPACE" | head -40

  2. Update the application version by changing the application_version input in the RAD platform and applying it via Update; a new image builds and the Deployment rolls out. authentik migrates its own schema on startup — guarded by a PostgreSQL advisory lock so concurrent pods don't collide — and the startup probe holds the rollout until migrations finish. Note that application_version = "latest" is pinned to a known-good release at build time (authentik publishes no latest tag).

  3. Inspect the co-located worker. ak worker runs in the same container as the server (started in the background by the cloud entrypoint), so its output is interleaved in the pod logs:

    POD=$(kubectl get pods -n "$NAMESPACE" -l app -o name | grep authentik | head -1)
    kubectl logs -n "$NAMESPACE" "$POD" --tail=200 | grep -iE 'worker|task'

    You should see worker startup and periodic task execution lines. Keep min_instance_count ≥ 1 so the worker is always processing.

  4. Manage secrets (never rotate the secret key — it invalidates all sessions and breaks encrypted fields):

    gcloud secrets list --project="$PROJECT" --filter="name~authentik"

  5. Open a database session for inspection or maintenance. The database and user names are tenant-prefixed — take them from the deployment outputs (database_name, database_user):

    INSTANCE=$(gcloud sql instances list --project="$PROJECT" --format="value(name)" --limit=1)
    gcloud sql connect "$INSTANCE" --user=<database_user> --database=<database_name> --project="$PROJECT"
    # e.g. inspect authentik's tables:  \dt authentik_*

Task 4 — Observe: Logging & Monitoring [Manual]​

  1. Logs — from the CLI or the Logs Explorer:

    gcloud logging read 'resource.type="k8s_container" AND resource.labels.namespace_name="'"$NAMESPACE"'"' \
      --project="$PROJECT" --limit=50

    Server and worker lines share the same stream; authentik logs structured JSON with an event field.

  2. Monitoring — open Kubernetes Engine → Workloads → the authentik workload for CPU/memory utilisation and restart counts, and the Cloud SQL instance dashboard for connections and query load. Review Alerting → Policies, and if you enabled the uptime check, confirm it is green under Monitoring → Uptime checks.

Task 5 — Troubleshoot & debug [Manual]​

Durable techniques for the failure modes you are most likely to hit. These are platform-level diagnostics and do not change with authentik releases.

  • Pod stays unready on first boot — usually just migrations. The first boot runs authentik's full Django migration suite; /-/health/ready/ returns 503 until it finishes. The startup probe budget is ~11 minutes (60 s delay + 40 × 15 s). Watch progress rather than assuming failure: 
    kubectl get pods -n "$NAMESPACE"
    kubectl logs -n "$NAMESPACE" "$POD" --tail=100
    # look for "Applying migration ..." lines, then the server/worker start banner
  • Database connection errors: confirm the Cloud SQL instance is RUNNABLE, the DB password secret exists, and the db-init Job completed. The entrypoint logs the resolved connection settings at startup (AUTHENTIK_POSTGRESQL__HOST/NAME/USER/SSLMODE): 
    gcloud sql instances list --project="$PROJECT"
    kubectl get jobs -n "$NAMESPACE"
    kubectl logs -n "$NAMESPACE" job/<db-init-job-name>
  • Blank page but the health endpoints return 200 — check which process owns :9000. The co-located ak worker also starts an HTTP listener; if it binds 0.0.0.0:9000 before the server, it answers every route — health endpoints included — with empty 200s (blank UI, phantom-healthy probes). The entrypoint now pins the worker to loopback ports (127.0.0.1:9001/9444/9301) so the server owns :9000. Verify with a body check: 
    curl -sk "$SERVICE_URL/" | wc -c   # non-zero = server answered; 0 = the worker won the bind
  • Image build failed: review Cloud Build history for the failed build's log. A MANIFEST_UNKNOWN / 404 pulling ghcr.io/goauthentik/server:<tag> means the requested application_version tag does not exist upstream — use a real release tag (or latest, which the module pins to a known-good release).
  • Pod events and scheduling issues: 
    kubectl describe pod -n "$NAMESPACE" "$POD" | tail -30
    kubectl get events -n "$NAMESPACE" --sort-by=.lastTimestamp | tail -20
  • 403 / permission errors: verify the workload service account's IAM roles and Workload Identity binding.

See the Configuration Guide's Configuration Pitfalls section for setting-specific gotchas (including the critical rule never to rotate AUTHENTIK_SECRET_KEY, and the SecretSync rule that synced-secret keys must not contain __).

Task 6 — Tear down [Automated]​

On the Deployments page, open the deployment and click the Trash icon (Delete). Delete runs terraform destroy and is irreversible (the deployment record is retained for history). If a deployment is stuck and the RAD platform can no longer manage it (for example after manual changes that conflict with the Terraform state), use Purge instead — it removes the deployment from RAD's records without destroying the cloud resources (it makes RAD forget the project). This removes everything the module created — the Kubernetes namespace and workload, Cloud SQL database, Secret Manager secrets, the media bucket, and Artifact Registry images. Resources owned by Services_GCP (the VPC, GKE cluster, shared Cloud SQL, registry) are managed separately and are not removed here.

Summary​

TaskTypeOutcome
1 — DeployAutomatedModule provisions the GKE workload, Cloud SQL (PostgreSQL 15), secrets, the GCS media bucket, builds the image, and runs DB init
2 — Access & verifyManual/-/health/ready/ returns 200 and the UI body is non-empty; log in as akadmin with the bootstrap password; create a first OIDC application + provider
3 — OperateManualInspect pods, update version (self-migrating), watch the co-located worker, manage secrets, DB access
4 — ObserveManualQuery Cloud Logging; review workload and Cloud SQL metrics
5 — TroubleshootManualDiagnose first-boot migration waits, database, init-job, build, and IAM issues
6 — Tear downAutomatedDelete (Trash) removes all module resources

By · Google Cloud certified: ACE · PCA · PCD · PDE · PSE · CDL

Last updated on