What problem does it solve? Deploying a new version to all users at once risks widespread outages when bugs slip through testing. This Skill implements a canary release strategy that gradually shifts traffic (1% → 5% → 25% → 100%) with automated metric gates, so issues affect only a small user slice before triggering rollback. ## Core Features & Use Cases - Argo Rollouts Configuration: Generates Kubernetes Rollout manifests with Istio traffic routing, staged weight steps, and Prometheus-based AnalysisTemplates for error rate, p99 latency, and success rate gates. - Nginx + Lua Alternative: Provides cookie-based sticky canary routing with an admin endpoint to adjust traffic weights when Istio is unavailable. - Automated Promotion Script: A Python CanaryManager that soaks each stage, queries Prometheus gates, and rolls back to 0% automatically on threshold breach. - Use Case: A team deploying a new payment service version routes 5% of traffic to the canary, verifies error rate stays under 1% for 10 minutes, then progressively promotes to 100% with automatic rollback if p99 latency exceeds 500ms. ## Quick Start Set up a canary release for my Kubernetes service using Argo Rollouts with Prometheus metric gates for error rate and latency.
