Learn more about Schema

Schema is a knowledge graph for how to operate infrastructure technology. Install the Schema MCP Server into the LLMs you use to develop and maintain applications. It gives you context on how telemetry and metrics interact, how to make multiple products work well together, and guidance on immediate actions to make your infra run better.

What is Schema?

A curated infrastructure knowledge graph — built from official documentation, incident postmortems, vendor guides, and practitioner expertise.

Sources
Official docsSource codeBenchmarks3rd party guidesO11y platformsOTel standards
SchemaSchemaInfrastructure knowledge graph

How do I use it?

Connect Schema to the LLM tools you already use via the Model Context Protocol (MCP).

Your LLM tools
Claude CodeCopilotCodexDatadog Bits AIDynatrace Davis
Schema MCP ServerInstall guide →

Why should I use it?

Schema gives LLMs the operational expertise they can't reliably get from training data.

Telemetry × config interactions
Understand how metrics, parameters, and configuration interact across your stack.
Multi-product operations
Know how infrastructure products operate best together — not just in isolation.
Grounded in original sources
Every recommendation links back to official documentation, vendor guides, and real incidents.
Calibrated thresholds
Metric ranges derived from production incidents, not generic advice.
Token efficient
Structured to minimize context window usage while maximizing information density.
Isolated and secure
No data ever leaves your boundary. Schema runs entirely within your environment.

What use cases is Schema for?

Explore operational scenarios where Schema adds the most value.

Cross-Stack Performance Tuning

Latency and throughput degradation caused by interacting services and configurations.

e.g.What are the top contributors to p99 latency including affected endpoints and downstream dependency timing?

Cost & Efficiency Under Real Workload

Waste or overspend that becomes visible only when utilization, architecture, and configuration are analyzed together.

e.g.Which resources are significantly over-provisioned relative to peak utilization?

Reliability Topology & Dependency Risk

Redundancy, failover, and failure propagation across interacting systems.

e.g.What is the blast radius if this dependency becomes unavailable?

Capacity & Bottleneck Propagation

How growth or pressure in one layer creates saturation in another.

e.g.Which services will exhaust a resource within N days at current growth?

Telemetry Semantics & Coverage

Whether the telemetry layer is good enough to support proactive cross-system reasoning at all.

e.g.Which services have gaps in metric coverage?

Change Impact & Configuration Drift

How changes in one layer alter behavior in adjacent layers — and whether those effects are visible.

e.g.Which recent deployments or config changes correlate with shifts in latency, saturation, or error rate?

Dependency Hygiene & Scaling Alignment

Whether interacting systems are configured and versioned in ways that scale safely together.

e.g.Which services are approaching dependency-imposed scaling limits before their own compute limits?

Toil-Causing Architectural Misalignment

Poorly aligned systems that require repeated operator intervention — not all toil is procedural.

e.g.Which services require repeated manual tuning due to misaligned autoscaling or resource limits?