BGswitch 🚀

Progressive Deployment Operator for Kubernetes

BGswitch is a Kubernetes operator that automates safe, progressive deployments with automatic traffic shifting, health monitoring, and rollback capabilities. Deploy with confidence using canary analysis and blue-green strategies.

🎯 Problem

Kubernetes rolling updates are risky:

• No gradual traffic shifting
• Limited health validation
• Manual rollback process
• No metric-based decision making

Result: Failed deployments cause downtime and impact users.

✨ Solution

BGswitch automates progressive deployments:

• 🔄 Gradual traffic shifting (10% → 25% → 50% → 100%)
• 📊 Metric-based health validation (Prometheus integration)
• ⚡ Automatic rollback on failures
• 🎛️ Manual promotion controls
• 📈 Real-time deployment status

🚀 Quick Start

Installation

# Install the operator
kubectl apply -f https://raw.githubusercontent.com/ghanatava/bgswitch/main/deploy/operator.yaml

# Verify installation
kubectl get pods -n bgswitch-system

Deploy Your First Progressive Deployment

apiVersion: bgswitch.dev/v1alpha1
kind: ProgressiveDeployment
metadata:
  name: my-app
spec:
  targetDeployment: my-app
  canarySteps: [10, 25, 50, 100]
  stepDuration: 5m
  metrics:
    errorRate:
      threshold: 1.0
      query: "rate(http_requests_errors_total{app='my-app'}[5m])"
    latency:
      threshold: 500
      query: "histogram_quantile(0.95, http_request_duration_seconds{app='my-app'})"
  autoPromote: true

kubectl apply -f my-progressive-deployment.yaml

# Watch progress
kubectl get progressivedeployment my-app -w

📊 How It Works

Stable (v1.0)     [████████████] 100%
                        ↓
Step 1 - Analyze  [███████████░] 90% stable, 10% canary
                        ↓ Metrics OK?
Step 2 - Promote  [████████░░░░] 75% stable, 25% canary
                        ↓ Metrics OK?
Step 3 - Promote  [█████░░░░░░░] 50% stable, 50% canary
                        ↓ Metrics OK?
Step 4 - Complete [░░░░░░░░░░░░] 100% canary → new stable

If metrics degrade at any step:

[██████░░] Rollback initiated → [████████████] Stable restored

🎛️ Features

Progressive Traffic Shifting

• Define custom canary steps (e.g., 5%, 10%, 25%, 50%, 100%)
• Configurable duration per step
• Replica-based traffic distribution

Health Monitoring

• Prometheus metric integration
• Custom PromQL queries
• Error rate tracking
• Latency monitoring
• Custom metric thresholds

Automatic Rollback

• Detects metric degradation
• Instant rollback to stable version
• Preserves original deployment
• Detailed rollback reasons

Manual Controls

# Promote to next step
kubectl bgswitch promote my-app

# Rollback immediately
kubectl bgswitch rollback my-app

# Pause progression
kubectl bgswitch pause my-app

# Get status
kubectl bgswitch status my-app

📖 Documentation

• Architecture
• Installation Guide
• Configuration Reference
• Metrics Guide
• Examples

🛠️ Development

Prerequisites

• Go 1.21+
• Docker
• kubectl
• kind (for local testing)

Build from Source

# Clone repository
git clone https://github.com/ghanatava/bgswitch.git
cd bgswitch

# Install dependencies
make install

# Run locally
make run

# Build operator image
make docker-build

# Run tests
make test

Local Testing

# Create local cluster
make cluster-up

# Install operator
make deploy

# Run example
kubectl apply -f examples/demo-app/

# Cleanup
make cluster-down

🎯 Roadmap

v0.1.0 (Current)

• Basic canary deployments
• Replica-based traffic shifting
• Prometheus integration
• Auto-rollback

v0.2.0 (Planned)

• Istio/Service Mesh integration for precise traffic control
• Webhook-based notifications (Slack, Discord)
• Advanced metric analysis (statistical tests)
• Multi-metric weighted decisions

v0.3.0 (Future)

• Blue-Green deployment strategy
• A/B testing support
• Custom webhook integrations
• Dashboard UI

🙏 Acknowledgments

Inspired by:

• Flagger - Progressive delivery toolkit
• Argo Rollouts - Advanced deployment strategies
• Spinnaker - Continuous delivery platform

📬 Contact

• GitHub Issues: Report bugs or request features
• Discussions: Ask questions

---

Built with ❤️ by [Ghanatava]

Making Kubernetes deployments safe and simple.