# Canary

> Implements gradual traffic shifting, automated canary deployments with metric analysis, rollback automation, and observability dashboards for progressive releases.

- Skill: `neuralblitz/canary` (Agent Skill)
- Install (CLI): `npx skillmds@latest add neuralblitz/canary`
- Raw SKILL.md: https://api.skillmd.com/api/skills/neuralblitz/canary/raw
- Safety review: PASS (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: DevOps & Infra, AI & ML, Deployment & Release, Monitoring & Observability
- Tags: Canary Deployment, Ci Cd, Flagger, Kubernetes, Observability, Rollback, Traffic Shifting
- License: MIT
- Author: NeuralBlitz (https://skillmd.com/u/neuralblitz)
- Updated: 2026-08-22
- Page: https://skillmd.com/skills/neuralblitz/canary

---


## What I do

- Implement gradual traffic shifting strategies
- Create automated canary deployments with metric analysis
- Design rollback automation based on health checks
- Build observability dashboards for canary analysis
- Integrate canary releases with CI/CD pipelines
- Configure intelligent traffic routing based on user segments

## When to use me

- When introducing new features to production
- When validating infrastructure changes
- When testing performance under real load
- When deploying to multiple regions progressively
- When working with A/B testing alongside deployments
- When reducing blast radius of potential issues

## Key Concepts

### Flagger Canary CRD

```yaml
apiVersion: flagger.app/v1alpha3
kind: Canary
metadata:
  name: myapp
spec:
  targetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: myapp
  service:
    port: 80
  canaryAnalysis:
    interval: 1m
    threshold: 10
    maxWeight: 50
    stepWeight: 10
    metrics:
      - name: request-success-rate
        threshold: 99
        interval: 1m
      - name: request-duration
        threshold: 500
        interval: 1m
```

### Deployment Strategy

- Start with small percentage (1-5%)
- Increase gradually if metrics stay healthy
- Stop and rollback if errors spike
- Analyze at each stage before proceeding
- Full rollout typically takes 30-60 minutes

### Analysis Criteria

- Error rate within acceptable threshold
- Response time not degraded
- No memory leaks or resource exhaustion
- Business metrics not negatively impacted
- No adverse effects on dependent services