CI/CD & DevOps11 min readPublished: 2026-09-08
Zero-Downtime Deployments: Implementing Canary and Blue-Green Strategies
Comprehensive guide to eliminating deployment downtime: traffic shifting at the ingress layer, stateful schema changes, and automated health verification.
CPL
CloudPlatformLabs Engineering
Platform Architecture Team
1. Blue-Green Deployments Explained
In Blue-Green deployment architecture, two identical production environments exist side by side. 'Blue' runs current live traffic, while 'Green' hosts the newly deployed version. Full smoke and regression tests execute against Green in isolation.
Once tests pass, routing is switched instantaneously at the load balancer or ingress controller. If unforeseen production bugs appear, traffic can be diverted back to Blue in seconds without redeploying code.
2. Traffic Splitting at the Ingress Controller
Using NGINX Ingress or AWS Application Load Balancer annotations, you can route a tiny percentage of live user requests (e.g. 5%) to the new version (Canary) based on headers, cookies, or weighted routing.
ingress-canary.yaml
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
name: order-service-canary
annotations:
nginx.ingress.kubernetes.io/canary: "true"
nginx.ingress.kubernetes.io/canary-weight: "10"
spec:
rules:
- host: api.cloudplatformlabs.dev
http:
paths:
- path: /orders
pathType: Prefix
backend:
service:
name: order-service-v2
port:
number: 80Related Topics & Tags
DevOpsZero DowntimeCanary DeploymentBlue GreenKubernetes
CPL
Written by CloudPlatformLabs Engineering
Platform Architecture Team
CloudPlatformLabs publishes independent, peer-reviewed cloud and DevOps architecture blueprints, hands-on tutorials, and infrastructure case studies. Our mission is to make advanced infrastructure engineering transparent, resilient, and developer-friendly.