Beyond services: your entire system mapped
Knowledge Graph provides three core capabilities:
| Capability | What it does | Problems solved |
|---|---|---|
| Entity Catalog | All entities and relationships | Unified view of services, hosts, pods, databases |
| Insights | Automatic issue detection | Problems surface before users report them |
| RCA Workbench | Timeline-based investigation | Guided troubleshooting with AI assistance |
Entity types discovered
| Type | Examples |
|---|---|
| Services | checkout-api, payment-service |
| Infrastructure | Hosts, nodes, pods, containers |
| Data | Databases, caches, message queues |
| Groupings | Namespaces, clusters, environments |
Insight types
| Type | What it detects |
|---|---|
| Anomalies | Traffic drops, latency spikes |
| Failures | 5xx errors, pod crashes |
| Saturation | Resources approaching limits |
| Amends | Deployments, config changes, scale events |
Questions answered
| With Knowledge Graph, you can answer… |
|---|
| What services run on this failing host? |
| What changed right before this issue started? |
| Which related entities are also having problems? |
| What’s the blast radius if this component fails? |
Script
Application Observability shows you services. Knowledge Graph goes deeper. It maps your entire system and automatically detects problems.
The Entity Catalog shows every entity in your environment: services, hosts, pods, databases, containers. It builds relationships automatically from your telemetry. What services run on this host? What databases does this service connect to? Knowledge Graph knows.
Insights are automatic issue detection. Anomalies, failures, errors, saturation, and changes, all surfaced without you configuring anything. Instead of checking dashboards hoping to find problems, problems come to you.
When you need to investigate, RCA Workbench gives you a timeline-based view correlating metrics, logs, traces, and insights. Navigate through related entities using the mind map. Grafana Assistant can even suggest hypotheses about root cause.
This shifts troubleshooting from “where do I start?” to “here’s the problem, here’s what changed, here are related issues.”