Skip to main content

Certification track: AI Tooling

RAGFlow on GKE Autopilot — Lab Guide

📖 Configuration Guide

Overview​

Estimated time: 45–90 minutes

RAGFlow is an open-source document intelligence and Retrieval-Augmented Generation (RAG) platform. It ingests PDFs, Word documents, HTML pages, and other formats, chunks and embeds them, stores vectors in Elasticsearch, and exposes a REST API and a web UI for knowledge base management and enterprise search. This lab takes you through the full operational lifecycle of the RAGFlow 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 RAGFlow 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 and access the running workload.
  • Perform day-2 operations — inspect, scale, update, and manage secrets and storage.
  • 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).
  • Elasticsearch_GKE deployed and its elasticsearch_endpoint output available — this is a hard deployment prerequisite; the plan is rejected if elasticsearch_hosts is empty when deploy_application = true.
  • 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 RAGFlow (GKE), set project_id and elasticsearch_hosts (the elasticsearch_endpoint output from your Elasticsearch_GKE deployment), and review the remaining 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 deploys the workload into the GKE Autopilot cluster, provisions a Cloud SQL (MySQL 8.0) database with its Secret Manager secrets, a Filestore (NFS) share for shared document storage, a Cloud Storage bucket for document artifacts, builds the container image, and runs a one-shot database-initialisation job. First deploys take roughly 20–35 minutes (Cloud SQL creation dominates).

  3. Connect to the cluster and discover the namespace 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"

    NS=$(kubectl get ns -o name | grep ragflow | head -1 | cut -d/ -f2)
    echo "Cluster: $CLUSTER   Namespace: $NS"
    kubectl get all -n "$NS"

Task 2 — Access & verify [Manual]​

  1. Confirm the workload is running and find its external address:

    kubectl get pods,svc -n "$NS"
    EXTERNAL_IP=$(kubectl get svc -n "$NS" \
      -o jsonpath='{.items[?(@.spec.type=="LoadBalancer")].status.loadBalancer.ingress[0].ip}')
    echo "External IP: $EXTERNAL_IP"
    curl -s "http://${EXTERNAL_IP}/v1/health"   # expect HTTP 200 with {"code":0}

    RAGFlow loads embedding models on first boot; if the health check does not return immediately, wait a few minutes for startup to complete.

  2. Open http://${EXTERNAL_IP} in a browser. On first visit RAGFlow presents a registration page — create an admin account with an email and password of your choice and sign in. There is no pre-provisioned admin credential in Secret Manager; the database password secret is for the MySQL backend only.

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

  1. Inspect the workload — deployment, pods, and (if enabled) the horizontal autoscaler and persistent volumes:

    kubectl get deploy,pods,hpa,pvc -n "$NS"
    kubectl describe deploy -n "$NS"

  2. Scale by changing the min/max instance inputs and clicking Update on the deployment details page — the module owns the workload spec, so scaling is a configuration change, not a manual kubectl scale (a manual edit would be reverted on the next apply). Note that min_instance_count is hard-capped at 1; scale-to-zero is not supported because RAGFlow loads embedding models at startup.

  3. Update the application version by changing the version input via Update on the deployment details page; a new image builds and a rolling update replaces the pods.

  4. Manage secrets, storage, and jobs:

    kubectl get secrets -n "$NS"
    gcloud secrets list --project="$PROJECT" --filter="name~ragflow"
    gcloud storage buckets list --project="$PROJECT"
    kubectl get jobs -n "$NS"          # DB-init and any scheduled jobs

  5. Open a database session for inspection or maintenance:

    INSTANCE=$(gcloud sql instances list --project="$PROJECT" --format="value(name)" --limit=1)
    gcloud sql connect "$INSTANCE" --user=ragflow --project="$PROJECT"

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

  1. Logs — from kubectl or the Logs Explorer:

    kubectl logs -n "$NS" deploy/"$(kubectl get deploy -n "$NS" -o jsonpath='{.items[0].metadata.name}')" --tail=50

    Logs Explorer filter: resource.type="k8s_container" AND resource.labels.namespace_name="<namespace>".

  2. Monitoring — open the GKE / Kubernetes dashboards and review pod CPU and memory utilisation, restart counts, and request metrics. The module also provisions an uptime check (when enabled); review Monitoring → Uptime checks and Alerting → Policies.

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 RAGFlow releases.

  • Pod not Ready / CrashLoopBackOff: inspect events and logs: 
    kubectl describe pod -n "$NS" <pod>          # Events section shows scheduling/probe/mount errors
    kubectl logs -n "$NS" <pod> --previous       # logs from the crashed container
  • Database connection errors: confirm the Cloud SQL instance is RUNNABLE, the DB password secret materialised into the namespace, and the init job completed. RAGFlow requires MySQL 8.0 — verify database_type = MYSQL_8_0.
  • Initialisation job failed: inspect the job and its pod logs: 
    kubectl get jobs -n "$NS"
    kubectl logs -n "$NS" job/<job-name>
  • Documents not processing (stuck in queue): confirm Redis is reachable — without Redis, RAGFlow's document workers never start. On GKE, an empty redis_host falls back to the NFS server IP; confirm the NFS instance is healthy.
  • Elasticsearch errors / failed indexing: verify elasticsearch_hosts points to the correct Elasticsearch endpoint and the cluster health is green.
  • Pending pod / no external IP: check kubectl describe pod events for resource or quota issues, and confirm the LoadBalancer Service has an assigned IP.
  • Image pull errors: confirm the image exists in Artifact Registry and the node service account can pull it.

See the Configuration Guide's Configuration Pitfalls section for setting-specific gotchas.

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 workload and namespace, Cloud SQL (MySQL) database, Secret Manager secrets, Cloud Storage bucket, Filestore (NFS) instance, static IP, 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. The Elasticsearch_GKE deployment must also be torn down separately.

Summary​

TaskTypeOutcome
1 — DeployAutomatedModule deploys the GKE workload, Cloud SQL (MySQL), NFS, Storage, secrets, and runs DB init
2 — Access & verifyManualConnect to the cluster; health check passes; register admin account and sign in
3 — OperateManualInspect workload, scale, update version, manage secrets/storage/jobs, DB access
4 — ObserveManualQuery Cloud Logging; review Cloud Monitoring metrics and uptime check
5 — TroubleshootManualDiagnose pod, database, init-job, Redis/Elasticsearch, scheduling, and image-pull issues
6 — Tear downAutomatedDelete (Trash) removes all module resources

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

Last updated on