Skip to main content

Observability Setup

This page explains how to set up Prometheus, Grafana, and distributed tracing for an llm-d deployment. All guides reference this page — set this up once and it works across every guide.

note

Commands in this page use ${NAMESPACE} for the namespace where your llm-d workload runs. Set it before following along:

export NAMESPACE=<your-llm-d-namespace>

Step 1: Install Prometheus and Grafana​

Skip this step if you already have Prometheus running in your cluster.

# Install Prometheus + Grafana into the llm-d-monitoring namespace
./guides/recipes/observability/install-prometheus-grafana.sh

For HTTPS/TLS (required by autoscalers like WVA):

./guides/recipes/observability/install-prometheus-grafana.sh --enable-tls

Verify the installation:

kubectl get pods -n llm-d-monitoring

Expected output:

NAME                                                     READY   STATUS    RESTARTS   AGE
alertmanager-llmd-kube-prometheus-stack-alertmanager-0    2/2     Running   0          30s
llmd-grafana-xxxxxxxxx-xxxxx                             3/3     Running   0          30s
prometheus-llmd-kube-prometheus-stack-prometheus-0        2/2     Running   0          30s

Platform-specific notes​

OpenShift​

OpenShift provides a built-in Prometheus stack via User Workload Monitoring. Enable it instead of installing a separate Prometheus:

GKE​

GKE clusters include Google Managed Prometheus (GMP) by default. GKE also provides a built-in inference gateway dashboard.

To use GMP as a Grafana data source, follow the GMP Grafana integration guide.

Manual ServiceMonitor Setup (Fallback)​

note

The recommended path is the llm-d helm charts, which create ServiceMonitors automatically when you include monitoring.values.yaml. Skip this section if you deployed that way.

Use this manual setup only as a fallback for workloads deployed outside the llm-d helm charts — e.g. CRD, KServe, or RHAII — where ServiceMonitors are not created for you. In that case, create them manually as shown below.

Find Your Service Labels​

ServiceMonitor selectors MUST exactly match your service labels:

# Find your EPP/Router service
kubectl get svc -n ${NAMESPACE} --show-labels

# View full labels for a specific service
kubectl get svc -n ${NAMESPACE} <epp-service-name> -o yaml

Note the exact app.kubernetes.io/component and app.kubernetes.io/name values.

Create ServiceMonitors​

For EPP:

apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
  name: <epp-servicemonitor-name>
  namespace: ${NAMESPACE}
spec:
  selector:
    matchLabels:
      app.kubernetes.io/component: <epp-component>
      app.kubernetes.io/name: <epp-name>
  namespaceSelector:
    matchNames:
      - ${NAMESPACE}
  endpoints:
    - port: metrics
      path: /metrics
      interval: 10s
      scheme: http
      bearerTokenFile: /var/run/secrets/kubernetes.io/serviceaccount/token

For vLLM:

apiVersion: monitoring.coreos.com/v1
kind: ServiceMonitor
metadata:
  name: <vllm-servicemonitor-name>
  namespace: ${NAMESPACE}
spec:
  selector:
    matchLabels:
      app.kubernetes.io/component: <vllm-component>
      app.kubernetes.io/name: <vllm-name>
  namespaceSelector:
    matchNames:
      - ${NAMESPACE}
  endpoints:
    - port: https
      path: /metrics
      interval: 10s
      scheme: https
      tlsConfig:
        insecureSkipVerify: true

Apply:

kubectl apply -f servicemonitor-epp.yaml
kubectl apply -f servicemonitor-vllm.yaml

Verify​

kubectl port-forward -n llm-d-monitoring svc/llmd-kube-prometheus-stack-prometheus 9090:9090
curl -sk 'https://localhost:9090/api/v1/targets' | jq '.data.activeTargets[] | select(.labels.namespace=="'${NAMESPACE}'") | {job: .labels.job, health: .health}'

Targets should show "health": "up". Then proceed to Step 2 for dashboards.

Step 2: Load Grafana Dashboards​

./guides/recipes/observability/load-llm-d-dashboards.sh

Verify dashboards were imported:

kubectl get configmaps -n llm-d-monitoring -l grafana_dashboard=1

Then access Grafana:

kubectl port-forward -n llm-d-monitoring svc/llmd-grafana 3000:80
# Open http://localhost:3000  (login: admin / admin)

Available dashboards:

DashboardWhat it shows
llm-d-vllm-overviewGeneral vLLM metrics overview
llm-d-failure-saturation-dashboardKey failure and saturation indicators
llm-d-diagnostic-drilldown-dashboardDetailed diagnostic metrics for troubleshooting
llm-d-performance-kv-cacheKV cache utilization and performance
llm-d-pd-coordinator-metricsPrefill/decode disaggregation metrics

Step 3: Install Distributed Tracing (Optional)​

Deploy the OTel Collector and Jaeger into the same namespace as your llm-d workload:

./guides/recipes/observability/install-otel-collector-jaeger.sh -n ${NAMESPACE}

Then access the Jaeger UI:

kubectl port-forward -n ${NAMESPACE} svc/jaeger-collector 16686:16686
# Open http://localhost:16686

For full tracing configuration across vLLM, the routing proxy, and the EPP, see Distributed Tracing.

Cleanup​

# Remove Prometheus and Grafana
./guides/recipes/observability/install-prometheus-grafana.sh -u -n llm-d-monitoring

# Remove OTel Collector and Jaeger
./guides/recipes/observability/install-otel-collector-jaeger.sh -u -n ${NAMESPACE}

Troubleshooting​

Autoscaler reports "http: server gave HTTP response to HTTPS client"​

The autoscaler is configured for HTTPS but Prometheus is serving HTTP. Enable TLS:

./guides/recipes/observability/install-prometheus-grafana.sh -u
./guides/recipes/observability/install-prometheus-grafana.sh --enable-tls

Metrics not appearing in Prometheus​

  1. Check that PodMonitors and ServiceMonitors exist:

    kubectl get podmonitors,servicemonitors -n ${NAMESPACE}

  2. Open http://localhost:9090/targets (after port-forwarding Prometheus) and check that vLLM and EPP targets show UP

  3. Confirm pods expose metrics:

    VLLM_POD=$(kubectl get pods -n ${NAMESPACE} -l app=my-model -o jsonpath='{.items[0].metadata.name}')
    kubectl port-forward -n ${NAMESPACE} ${VLLM_POD} 8000:8000
    curl http://localhost:8000/metrics | head -20

Grafana dashboards show "No data"​

  1. Verify the Grafana datasource points to the correct Prometheus URL
  2. Check that metrics are flowing in Prometheus first
  3. If using TLS, ensure the Grafana datasource is configured for HTTPS with the correct CA certificate