Causely MCP server

What it does

Causely exposes a causal model for reliability engineering through an MCP server. Agents query root cause analysis, blast radius, and service health directly instead of analyzing raw telemetry. For teams using AI agents to investigate incidents and understand system dependencies.

• Triage alerts by querying the causal model to identify root causes and blast radius before investigating raw metrics.
• Investigate Kubernetes cluster issues by correlating service dependencies and health status through structured tools.
• Generate postmortem summaries and health reports by extracting causal relationships and incident correlations from the model.

Tools this server exposes

Tool	Does
root_cause	Query Causely's causal model to determine the root cause of an issue.
blast_radius	Identify which services and systems are affected by a given failure.
service_health	Check the health status and dependencies of a service.
alert_triage	Route alerts through causal reasoning to prioritize investigation.
change_impact	Analyze the impact of changes on connected services.
correlated_incidents	Find incidents that share causal relationships.
kubernetes_investigation	Query Kubernetes cluster state with causal context.

Before you enable it

An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.