Observe by Snowflake announced general availability of a redesigned MCP server and a new CLI with full feature parity, giving AI agents and engineers direct programmatic access to telemetry data. The rearchitected MCP server drops the previous LLM intermediary, letting agents query Observe's APIs and context graph directly, reducing latency and cost. Both tools ship with prebuilt skills for common observability workflows like incident investigation, failure tracing, and outlier detection, and work with agent environments like Claude Code as well as interactive terminal sessions.
Questions this post answers
What changed in Observe by Snowflake's new MCP server compared to the old one?
The redesigned MCP server removes the custom LLM intermediary that previously interpreted queries and returned answers through a single endpoint. Agents now connect directly to the same APIs that power Observe's frontend, letting them query the context graph, explore data sets, and write OPAL queries themselves, cutting latency and cost for every call. Engineers weighing observability platforms for AI agent workflows can track these architecture shifts on daily.dev.
Does Observe by Snowflake have a CLI that supports AI agent workflows?
Yes, a new agent-compatible CLI was released with full parity to the MCP server, meaning every operation exposed to agents through MCP is also available as a CLI command. It works in agent environments like Claude Code as well as interactive terminal sessions, and ships with prebuilt skills for tasks like incident investigation and outlier detection. Teams automating incident response with AI agents can follow tooling updates like this via daily.dev.