Infrastructure evidence
Retain operational history and make past system state easier to query and verify.
Open-source cloud-native operations
Less operational noise. More important work.
nowake.ai builds open-source, AI-assisted tools that make infrastructure work easier to inspect, automate, and prioritize.
operational-noise.index inspectable
Alerts and notification noise
Triage should preserve context and lift meaningful cases.Manual maintenance
Repeated work should become bounded, reviewable automation.Access, safety, and audit
Production changes need least privilege and accountable records.Recovery and capacity work
Known procedures should be easier to verify and repeat.What we build
Tools for evidence, automation, access, and signal.
AI is part of the route, but not the product promise. The shared goal is practical infrastructure work that engineers can inspect, trust, and control.
Kubernetes-native automation
Reduce repeated infrastructure work while preserving Kubernetes patterns.
Access and auditability
Make temporary access, operation records, and production changes easier to explain.
Alert intelligence
Reduce notification noise and surface cases that deserve attention with history and feedback.
Active projects
Current work
kube-insight
Retained Kubernetes evidence for humans and agents: sanitized list/watch history, facts, topology, and read-only SQL, API, and MCP surfaces.
svc-lb-mux
Service LoadBalancer Multiplexer: a Kubernetes controller that lets many application-facing LoadBalancer Services share one provider-managed Layer 4 load balancer.
Principles
No magic in the control plane.
Open source first
Evidence before automation
Operator control before autonomy
Least privilege before convenience
Auditability before magic
Boring reliability over AI spectacle