Deployment Verification¶

After deploying Geospatial Studio, verify that all components are working correctly before proceeding to the workshop labs.

Verification Time

Estimated time: 10 minutes

Step 1: Check Pod Status¶

Verify all pods are running:

Local DeploymentCluster Deployment

export KUBECONFIG="$HOME/.lima/studio/copied-from-guest/kubeconfig.yaml"
kubectl get pods

kubectl get pods -n <your-namespace>

Expected output - all pods should show Running or Completed:

NAME                              READY   STATUS    RESTARTS   AGE
geofm-gateway-xxx                 1/1     Running   0          5m
geofm-ui-xxx                      1/1     Running   0          5m
geofm-mlflow-xxx                  1/1     Running   0          5m
geofm-geoserver-xxx               1/1     Running   0          5m
postgresql-xxx                    1/1     Running   0          8m
keycloak-xxx                      1/1     Running   0          8m
minio-xxx                         1/1     Running   0          8m
redis-xxx                         1/1     Running   0          8m

Troubleshooting

If any pods show Error, CrashLoopBackOff, or Pending, check the Troubleshooting Guide.

Step 2: Verify Service Access¶

Studio UI¶

Open https://localhost:4180 (local) or your cluster URL
Accept the security warning (self-signed certificate)
You should see the login page

Enter credentials:
Username: testuser
Password: testpass123
Click "Sign In"
You should see the Studio home page

Geospatial Studio Home Page - Main Interface

API Access¶

Test the API endpoint:

# For local deployment
curl -k https://localhost:4181/health

Expected response (authentication redirect):

<a href="http://keycloak.default.svc.cluster.local:8080/realms/geostudio/protocol/openid-connect/auth...">Found</a>

Authentication Required

The /health endpoint requires authentication and will redirect to Keycloak login. This confirms the API gateway is running and properly configured with authentication. To make authenticated API calls, you'll need to create an API key (see Step 3 below).

Alternative: Check Pod Status

To verify the API gateway is running without authentication, check the pod status:

kubectl get pods | grep geofm-gateway

The pod should show Running status.

Step 3: Create and Test API Key¶

Create API Key¶

On the Studio home page, click "Manage your API keys"
Click "Generate new key"
Copy the generated API key
Save it to a config file:

echo "GEOSTUDIO_API_KEY=<your-api-key>" > ~/.geostudio_config_file
echo "BASE_STUDIO_UI_URL=https://localhost:4180" >> ~/.geostudio_config_file

Test API Key¶

# Set environment variables
export STUDIO_API_KEY="<your-api-key>"
export UI_ROUTE_URL="https://localhost:4180"

# Test API call
curl -k -X GET "${UI_ROUTE_URL}/studio-gateway/v2/models" \
  -H "X-API-Key: ${STUDIO_API_KEY}"

Expected response: JSON array of models (may be empty initially)

Step 4: Verify Backend Services¶

GeoServer¶

Open http://localhost:3000/geoserver
Login with:
Username: admin
Password: geoserver
You should see the GeoServer admin interface

MLflow¶

Open http://localhost:5000
You should see the MLflow tracking UI
No login required

MinIO¶

Open https://localhost:9001
Login with:
Username: minioadmin
Password: minioadmin
You should see the MinIO console

Step 5: Onboard Sandbox Models¶

Onboard placeholder models for testing:

export STUDIO_API_KEY="<your-api-key>"
export UI_ROUTE_URL="https://localhost:4180"

./deployment-scripts/add-sandbox-models.sh

Expected output:

✅ Onboarding sandbox model: geofm-sandbox-models
✅ Model onboarded successfully

Verify in UI: 1. Navigate to "Models" page 2. You should see "geofm-sandbox-models" listed

Step 6: Test Python SDK¶

Create a test script:

# test_sdk.py
from geostudio import Client
import urllib3
urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)

# Initialize client
client = Client(geostudio_config_file=".geostudio_config_file")

# List models
models = client.list_models()
print(f"✅ Found {len(models)} models")

# List datasets
datasets = client.list_datasets()
print(f"✅ Found {len(datasets)} datasets")

print("✅ SDK verification complete!")

Run the test:

python test_sdk.py

Expected output:

✅ Found 1 models
✅ Found 0 datasets
✅ SDK verification complete!

Step 7: Verification Checklist¶

Confirm all items are working:

Troubleshooting Common Issues¶

Cannot Access UI¶

Connection refused or timeout

Local Deployment:

# Check port forwarding
ps aux | grep "kubectl port-forward"

# Restart port forwarding if needed
export OC_PROJECT=default
kubectl port-forward -n $OC_PROJECT deployment/geofm-ui 4180:4180 >> studio-pf.log 2>&1 &

Cluster Deployment:

# Check ingress/route
kubectl get ingress -n <namespace>
oc get routes -n <namespace>

API Key Not Working¶

401 Unauthorized error

Verify API key is correct
Check it's properly set in environment variable
Ensure no extra spaces or newlines
Try generating a new API key

Pods Not Running¶

Pods in Error or CrashLoopBackOff state

# Check pod logs
kubectl logs <pod-name>

# Check pod events
kubectl describe pod <pod-name>

# Common issues:
# - Insufficient resources
# - Image pull errors
# - Configuration errors

SDK Connection Issues¶

Cannot connect to Studio

Verify BASE_STUDIO_UI_URL is correct
Check API key is valid
For local deployment, ensure port forwarding is active

Try with verify_ssl=False for self-signed certificates:

client = Client(
    geostudio_config_file=".geostudio_config_file",
    verify_ssl=False
)

Performance Verification¶

Check Resource Usage¶

# Check node resources
kubectl top nodes

# Check pod resources
kubectl top pods -n <namespace>

Database Connection¶

# Test PostgreSQL connection
kubectl exec -it <postgresql-pod> -- psql -U postgres -c "SELECT version();"

Object Storage¶

# List MinIO buckets
kubectl exec -it <minio-pod> -- mc ls local/

Next Steps¶

✅ Verification Complete! Your Geospatial Studio deployment is ready.

Proceed to:

Introduction → - Learn about Geospatial Studio
Lab 1 → - Start the hands-on labs

Additional Resources¶

← Back to Deployment Next: Introduction →