Observe launches MCP server and CLI for AI agents to access telemetry

Observe by Snowflake has announced general availability of a redesigned MCP (model context protocol) server and a new CLI tool designed to give AI agents direct, programmatic access to observability data. The rollout is happening now across all clusters, with eu-2 and ca-1 regions following shortly.

The core innovation is architectural. The original MCP server contained a custom LLM harness: agents sent queries to a single endpoint, where Observe's LLM interpreted the question and returned an answer. This design limited agent visibility into Observe's data structure and restricted which data sets agents could explore—a limitation that was costly both for Observe and its customers. The new MCP server eliminates that LLM intermediary entirely. Agents now connect directly to Observe's platform APIs, gaining access to the same capabilities available in the Observe UI: APM service maps, OpenTelemetry data collection setup, and active alert listing. This direct access lets agents query Observe's context graph, determine which data sets are most valuable to query for accurate answers, and write efficient OPAL (Observe's query language) to run precise observability queries against any telemetry data.

The new CLI provides the same programmatic access from the command line, with full parity to the MCP server. It works in agent environments like Claude Code as well as interactive terminal sessions. Engineers can use it to compose, automate, and extend observability workflows. Some workflows run autonomously in the background, handling routine tasks; others are designed for interactive use, requiring an engineer to guide an investigation directly. Both the CLI and MCP server ship with a library of prebuilt skills—structured workflows based on how Observe engineers solve common observability tasks. These skills enable agents and engineers to investigate incidents, trace failures, validate changes, find n+1 issues, and detect outliers without additional setup after configuration.

In practice, engineering teams are already using these tools to build custom alert-triage agents that automatically query Observe when an alert fires and pull production telemetry for debugging context, and copilots that work alongside humans during incident investigations. The removal of the LLM intermediary makes this capability more cost-efficient to operate, reducing both latency and overhead for every call. Engineers can access the MCP server by logging into their tenant, navigating to "Manage account" → "MCP server details", and following setup instructions for their agent of choice.