Observe by Snowflake Launches MCP Server and CLI for AI Agents

Observe by Snowflake has announced the general availability of a redesigned MCP server and a new CLI, both of which enable AI agents to access production telemetry and observability data directly. This announcement addresses a new reality in incident response and debugging: AI agents are now primary consumers of operational data. Coding assistants can investigate errors before an engineer opens a ticket; AI SREs can correlate failures across services without human intervention. As agents become more sophisticated in reasoning about production systems, observability platforms must work for both humans and machines—through UIs, APIs, and interfaces that agents can call and automate.

The previous Observe CLI covered only a narrow slice of platform functionality, with most capabilities inaccessible from the terminal. The new CLI is agent-compatible and provides programmatic access to Observe's full capabilities—the same surface available through the MCP server and the UI. It works from agent environments like Claude Code as well as interactive terminal sessions. Engineers can compose workflows that run autonomously in the background (handling routine tasks without human presence) or interactively (with an engineer guiding the investigation). The CLI is designed as a programmatic interface for composing, automating, and extending observability workflows.

Both the CLI and MCP server ship with prebuilt skills: structured workflows built on how Observe engineers solve common observability tasks. Agents and engineers can use these skills to investigate incidents, trace failures, validate changes, find n+1 issues, and detect outliers. They work immediately after configuration, requiring no additional setup. The most significant change is architectural and economic. The original MCP server had a custom LLM harness built in: it received a user query, handled the reasoning, and returned an answer. In the new design, that intermediary is removed. Agents now connect directly to Observe's APIs, which gives them direct visibility into Observe's data structures and permission to explore data sets. This removes the latency and cost of the old LLM middleman. In the old model, agents had to send queries to a single endpoint where Observe's LLM interpreted the question; they lacked direct access to data structure and couldn't explore data sets autonomously—expensive for both Observe and its customers. The new MCP server gives agents access to the same APIs that power Observe's frontend, including APM service maps, OpenTelemetry data collection setup, and active alert listing. Agents can query Observe's context graph directly, determine which data sets are most valuable, and write efficient OPAL queries to run precise observability queries against any telemetry data. General availability is rolling out now across all clusters, with eu-2 and ca-1 regions to follow shortly.