AI SRE platform for Kubernetes built on an eBPF telemetry layer. Autonomously monitors services, spawns AI agents to root-cause issues using complete runtime context (traces, metrics, logs, K8s state, deployments, code changes), performs deployment verification, and generates fix pull requests. The eBPF layer provides the data; AI agents drive operational actions — investigations, verifications, fixes.
Metoro is a Kubernetes observability and AI SRE platform built on an eBPF telemetry layer. The eBPF DaemonSet captures traces, metrics, logs, and network flows at the kernel level without application instrumentation or sidecar injection.
AI agents use this telemetry to investigate failures: when a service degrades, the agent correlates the failure with the deployment or code change that caused it. For a defined set of failure classes (missing environment variables, misconfigured service references), Metoro generates and opens a pull request with the fix applied. Deployment verification monitors services during rollouts and flags regressions before full rollout completes.
Free tier for basic monitoring. Kubernetes-only; no VM or serverless coverage.
Key Features
eBPF telemetry collection: captures traces, metrics, logs, Kubernetes state, and network flows at the kernel level via a privileged DaemonSet — no application instrumentation or sidecar injection required
AI root-cause investigation: correlates service degradations with deployment events, code changes, and configuration modifications using the complete runtime context from the eBPF data layer
Automated fix PRs: for narrow failure classes (missing env vars, misconfigured references), generates and opens a pull request with the fix applied rather than surfacing a diagnosis only
Deployment verification: monitors key service metrics during rollouts and flags regressions automatically before the rollout completes — operates on eBPF telemetry without synthetic monitors
Zero-instrumentation coverage: services emit no code changes to produce traces and metrics — useful for legacy services or third-party containers where instrumentation is not feasible
Kubernetes-native correlation: reads deployment events, pod lifecycle, and config changes from the Kubernetes API alongside eBPF telemetry for combined root-cause analysis
