Asynkron.OtelMcp
Asynkron.OtelMcp combines an OpenTelemetry collector, time-series pipeline, and search interface with a Machine Control Protocol (MCP) surface so AI agents can reason about your telemetry. Build automated runbooks, anomaly investigations, and health dashboards that stay current as your distributed systems evolve.
Project links
- Source repository: asynkron/OtelMCP
Key capabilities
- Unified telemetry ingest – Ship traces, metrics, and logs from any OpenTelemetry SDK or collector and normalize them with zero-copy processing.
- AI-native query surface – Expose observability data through MCP so LLM-powered copilots can ask questions, correlate signals, and propose remediation steps.
- Scalable storage – Retain high-cardinality data by streaming into a columnar store optimized for actor workflows and multi-tenant workloads.
- Programmable workflows – Trigger automations when collectors detect new patterns, service regressions, or saturation events.
Documentation map
- Getting Started – Stand up the collector and import your first telemetry streams.
- MCP Agents – Wire up agent prompts, guardrails, and feedback loops.
- Operations – Scale storage tiers, retention policies, and alerting.
- Reference – Configuration schema, ports, and supported exporters.
Looking for concrete examples? Check the samples directory for docker-compose stacks and IaC modules you can drop into your environment.