Powering Agentic Observability with the Observe MCP Server

Observe is an AI-powered observability platform that enhances performance, troubleshooting, and cost efficiency by correlating logs, metrics, and traces within an open data lake. It aims to expand access to its capabilities through the Model Context Protocol (MCP) server, which connects AI agents with external resources, allowing them to gather context and generate insights, code changes, or error investigations. Unlike typical MCP implementations that simply wrap APIs, Observe has developed an AI-first approach by incorporating a Knowledge Graph, which indexes infrastructure and business data with enriched metadata to assist AI agents in generating contextually relevant queries. These queries, structured using a JSON schema, are then translated into the OPAL query language to provide complex insights, which are represented visually through markdown-based tables and PNGs. By hosting its MCP server, Observe offers controlled execution, remote configuration, and rapid updates, allowing enterprise customers to integrate both their private data and SaaS data for comprehensive observability.