Investigate Kubernetes Health and Processes with the Datadog MCP Server
Datadog, Monday, November 9th, 2026
Datadog adds Kubernetes control plane health tools, Live Processes search and a processes skill to its MCP Server for AI agents.
Datadog has expanded the infrastructure investigation capabilities of its MCP Server, which connects AI assistants to observability data.
Four read-only tools return health metrics for a named cluster's API server, scheduler, controller manager and etcd in one call, covering signals such as request error rates, scheduling durations, work queue depth and leader changes.
The tools take a cluster name and an optional time window that defaults to the last 10 minutes and can cover up to 4 hours.
Live Processes search gives agents process-level evidence, and a processes skill guides them to the right data source, with the aim of letting agents investigate issues without first searching the metric catalog.