Documentationbreadcrumb arrow Grafana Cloudbreadcrumb arrow Knowledge graphbreadcrumb arrow Get startedbreadcrumb arrow Prerequisites
Grafana Cloud

Prerequisites

The knowledge graph works with data you’re already sending to Grafana Cloud. You can activate the knowledge graph if you have enabled at least one of the following Grafana Cloud products:

Additional product requirements for specific entity types include:

  • Kubernetes infrastructure (K8s nodes, clusters, namespaces) → Kubernetes Monitoring
  • Bare metal infrastructure (Linux/Windows hosts) → Node Exporter or equivalent host metrics
  • Application frameworks (Spring Boot, Flask, Node.js, etc.) → Application Observability
  • Frontend applications (browser, SPA) → Frontend Observability
  • Cloud provider entities (AWS Lambda, RDS, ELB, etc.) → Cloud Service Provider Observability
  • Databases (PostgreSQL, MySQL, MongoDB, etc.) → Database Observability
  • Message queues and caches (Redis, Kafka, RabbitMQ, etc.) → Application Observability with technology-specific exporters

For information about which metrics the knowledge graph requires, refer to Required metrics and labels troubleshooting.

Supported technologies

The knowledge graph automatically discovers and monitors a wide range of infrastructure, applications, and services. The level of support varies:

  • Is Discovered: The knowledge graph automatically detects these entities from telemetry and displays them in the entity catalog with base RED metrics (Rate, Error, Duration)
  • Has Insights: The knowledge graph generates technology-specific insights (Saturation, Amend, Anomaly, Failure, Error) beyond base metrics

Compatibility matrix

TechnologyIs DiscoveredHas Insights
Infrastructure
K8s Linux nodes✓✓
K8s Windows nodes✓
Linux bare metal✓✓
Windows bare metal✓
Application Frameworks
Spring Boot / JVM✓✓
Flask✓
Node.js✓
Go✓
.NET✓
PHP-FPM✓
Loopback✓
Express✓
gRPC✓
Frontend Applications
Browser (Faro SDK)✓✓
Web Servers
NGINX Ingress✓✓
NGINX (standalone)✓✓
Traefik✓✓
Caddy✓
Tomcat✓
Databases
PostgreSQL✓✓
MySQL✓✓
MongoDB✓
Redis✓✓
Elasticsearch✓✓
OpenSearch✓
ClickHouse✓
Memcached✓
AWS Cloud Services
AWS EKS✓✓
AWS RDS✓✓
AWS Lambda✓
AWS DynamoDB✓
AWS S3✓
AWS SQS✓
AWS API Gateway✓
AWS Application ELB✓✓
AWS Network ELB✓✓
AWS EC2✓
AWS ECS✓
Azure Cloud Services
Azure AKS✓✓
GCP Cloud Services
GCP GKE✓✓
GCP CloudSQL✓
GCP Compute Engine✓
Service Mesh
Istio✓
Message Queues
Kafka✓✓
RabbitMQ✓✓
Monitoring
Prometheus✓✓
Alertmanager✓✓
Victoria Metrics✓
Grafana✓
Storage
Portworx✓
CI/CD
Jenkins✓
Flux✓✓
Security and Networking
Cert Manager✓✓
Blackbox Exporter✓✓

Discovery requirements by technology type

Infrastructure (Nodes, Clusters, Namespaces):

  • Product: Requires Kubernetes Monitoring or Node Exporter
  • Instrumentation:
    • Kubernetes infrastructure: Kubernetes Monitoring enabled
    • Bare metal: Node Exporter or equivalent host metrics

Application frameworks (Spring Boot, Flask, Node.js, etc.):

Frontend applications (Browser, SPA):

  • Product: Requires Frontend Observability
  • Instrumentation:
    • Grafana Faro Web SDK for browser telemetry
    • Automatic discovery via Frontend Observability settings
    • W3C Trace Context for backend correlation

Cloud services (AWS, Azure, GCP):

  • Product: Requires Cloud Service Provider Observability for cloud entity discovery
  • Instrumentation:
    • Cloud provider metadata in metric labels
    • Amazon CloudWatch, Azure Monitor, or GCP Cloud Monitoring integration

Databases (PostgreSQL, MySQL, MongoDB, etc.):

  • Product: Requires Database Observability for granular database integrations
  • Instrumentation:
    • Technology-specific exporters (for example, postgres_exporter, mysql_exporter)
    • Service mesh telemetry (optional, for automatic discovery)

Message queues and caches (Redis, Kafka, RabbitMQ, etc.):

  • Product: Requires Application Observability with technology-specific exporters
  • Instrumentation:
    • Technology-specific exporters (for example, redis_exporter, kafka_exporter)
    • Service mesh telemetry (optional, for automatic discovery)

Insight levels

  • Is Discovered: Technology appears in the entity catalog with base RED metrics and insights (Saturation, Amend, Anomaly, Failure, Error)
  • Has Insights: Technology receives technology-specific operational health alerts beyond generic monitoring. Examples include: Redis cluster health, Kafka replication status, Elasticsearch shard management, JVM garbage collection saturation.

For detailed instrumentation instructions, refer to Get started with the knowledge graph.

Infrastructure metrics

The knowledge graph scans your metrics to identify infrastructure and USE metrics for the entities it discovers.

Currently, the knowledge graph collects infrastructure metrics from Kubernetes environments. To learn how to set up monitoring for your Kubernetes cluster, refer to Configure Kubernetes Monitoring.

If you’ve set up Kubernetes monitoring but aren’t seeing entities and information in the knowledge graph, refer to Troubleshooting for assistance with common issues.

Grafana Labs is working to expand support for other metric sources and custom metrics. If you’re interested in integrating your custom telemetry with the knowledge graph, submit a support ticket.

RED metrics

The knowledge graph matches RED metrics (Request rate, Error rate, Duration) with the entities it detects through Kubernetes monitoring or Application Observability. By default, the knowledge graph supports Grafana Cloud Application Observability and Istio.

To collect RED metrics, you need to instrument your applications with OpenTelemetry and send the data to Grafana Cloud.

If you can’t see your OpenTelemetry instrumented services in the knowledge graph, refer to Troubleshooting for assistance with common issues.