You may have heard of the pillars of observability: metrics, logs, traces, and, depending on who you ask, profiles. As systems grow in complexity, correlating these signals for rapid incident detection, root cause analysis, and performance optimization still demands knowing specialized query languages and complex toolchains, even with OpenTelemetry.
Model Context Protocol (MCP) servers give assistants and tools a secure, standard way to work with your telemetry using natural language; this session introduces MCP and demonstrates exploring real observability data with Grafana MCP, while showing how the same approach applies to other MCP compatible tools or custom servers.
