When a Kubernetes cluster misbehaves at 3 am, the person on call is not always the one who understands it best. New Relic is betting that a specialised AI agent can narrow that gap, and has launched the New Relic Autopilot Kubernetes Agent to identify the cause of an incident and draft the code to fix it.

The observability company says the agent is aimed at operations teams, including site reliability engineers (SREs) and platform engineers. It is a domain-specific extension of New Relic Autopilot, the company’s fleet of SRE agents that investigate incidents and speed up response. Where the wider Autopilot handles full-stack observability queries, the Kubernetes agent concentrates on containerised environments, which New Relic describes as the setups where detection and resolution are traditionally hardest.

According to New Relic, the agent automatically pinpoints root causes and generates context-aware fix code that engineers review and run themselves. It runs more than 90 specialised diagnostic detectors across Kubernetes events, object descriptions, logs and application performance monitoring (APM) traces. The company says it is powered by New Relic Ground Truth, its context engine, which evaluates correlated operational data built on OpenTelemetry, the open standard for collecting telemetry. New Relic contrasts this with generic large language models, which it says offer speculative advice because they lack operational context.

The company frames the product around a skills problem. It cites the Cloud Native Computing Foundation (CNCF) as finding that 66% of organisations running generative AI rely on Kubernetes, and argues that the resulting spread of clusters is exposing an expertise gap. In its telling, Kubernetes shifts operations from managing fixed machines to managing a dynamic, self-healing distributed system, and diagnosing faults across large container estates has demanded deep knowledge that few teams have. During outages, it says, engineers can lose hours stitching together fragmented dashboards or sifting through generic AI advice.

Brian Emerson, New Relic’s chief product officer, said: “When critical Kubernetes infrastructure fails, engineers need clear answers, not just more alerts. The New Relic Autopilot Kubernetes Agent embeds specialized, agentic AI directly into the engineer’s containerized workflows to eliminate the guesswork from complex cluster incidents while keeping the engineer safely in control. Because it works natively across any telemetry source, it provides a level of full-stack observability that specialized or closed-system tools cannot match.”

The announcement, however, does not say whether the agent is generally available or in preview, what it costs, or which Kubernetes distributions and cloud providers it supports.