ReleaseBytes
snowflake Snowflake Blog ·

Observe by Snowflake Announces GA of Redesigned MCP Server and New CLI for AI Agents

blogobservabilitysnowflakegaengineer
feature announcement

Observe by Snowflake announces the general availability of its redesigned Model Context Protocol (MCP) server and a new Command Line Interface (CLI). These tools provide AI agents and engineers with programmatic access to telemetry data, enabling automation and direct interaction with the Observe platform. The update focuses on enhancing cost-efficiency and empowering agents to query production data directly, bypassing the UI for tasks like alert triage and incident investigation. It also introduces prebuilt skills to streamline common observability workflows.

  • →General Availability of Redesigned Observe MCP Server and New CLI
  • →Expanded Programmatic Access via Agent-Compatible CLI
  • →Prebuilt Skills for Common Observability Workflows
  • →Rearchitected MCP for Cost-Efficiency and Direct Agent Access
Features (3) ›
  • General Availability of Redesigned Observe MCP Server and New CLI

    Observe by Snowflake has announced the general availability of its rearchitected Model Context Protocol (MCP) server and a new Command Line Interface (CLI). These tools offer full parity, ensuring every operation available through the MCP server for agents is also accessible via CLI commands, rolling out now across clusters.

  • Expanded Programmatic Access via Agent-Compatible CLI

    The new Observe CLI provides comprehensive programmatic access to the platform's capabilities, matching the functionality available through the MCP server and UI. This enables engineers and AI agents to interact with Observe from various environments, automate routine tasks, and guide incident investigations.

  • Prebuilt Skills for Common Observability Workflows

    Both the new CLI and redesigned MCP server ship with a library of prebuilt skills, which are structured workflows based on Observe engineers' best practices. These skills allow agents and engineers to immediately investigate incidents, trace failures, validate changes, and detect outliers without additional setup.

Enhancements (1) ›
  • Rearchitected MCP for Cost-Efficiency and Direct Agent Access

    The redesigned MCP server removes the previous LLM intermediary, allowing agents to connect directly to Observe's APIs. This significantly reduces latency, overhead, and operational costs. It also provides agents with direct visibility into Observe's data structure, enabling more accurate queries and efficient use of OPAL.

Read the original announcement →

https://www.snowflake.com/content/snowflake-site/global/en/blog/observe-mcp-server-cli-ai-agents-telemetry

Related releases