# Blue-Green and Canary Deployments {#blue-green-and-canary-deployments}


Zero-downtime deployment strategies for Ever Gauzy.


## Blue-Green Deployment {#blue-green-deployment}






```mermaid
graph LR
    LB[Load Balancer] --> Blue[Blue v1.0]
    LB -.->|Switch| Green[Green v1.1]
```






### Process {#process}


1. **Blue** is the current production (v1.0)
2. **Green** is deployed with new version (v1.1)
3. Run smoke tests against Green
4. Switch load balancer to Green
5. Blue becomes standby for rollback


### Kubernetes Implementation {#kubernetes-implementation}


```yaml
# service.yaml
apiVersion: v1
kind: Service
metadata:
  name: gauzy-api
spec:
  selector:
    app: gauzy-api
    version: green # Switch between blue/green
  ports:
    - port: 3000
```


### Rollback {#rollback}


Simply switch the service selector back to `version: blue`.


## Canary Deployment {#canary-deployment}






```mermaid
graph LR
    LB[Load Balancer] -->|90%| Stable[Stable v1.0]
    LB -->|10%| Canary[Canary v1.1]
```






### Process {#process-1}


1. Deploy canary (v1.1) alongside stable (v1.0)
2. Route 10% of traffic to canary
3. Monitor error rates and latency
4. Gradually increase canary traffic
5. Full rollout or rollback


### Nginx Canary {#nginx-canary}


```nginx
upstream gauzy {
    server api-v1:3000 weight=9;
    server api-v2:3000 weight=1;
}
```


## Related Pages {#related-pages}


- [Production Deployment](https://sourcey-gauzy.pages.dev/devops/production-deployment/index.md) — deployment guide
- [CI/CD Pipeline](https://sourcey-gauzy.pages.dev/deployment/ci-cd/cicd-pipeline-guide/index.md) — CI/CD
- [Health Checks](https://sourcey-gauzy.pages.dev/observability/health-checks/index.md) — health monitoring