Canary deployments in Nginx - Mini Project: Build & Apply

Canary Deployments with Nginx

📖 Scenario: You are working as a DevOps engineer for a web service. Your team wants to test a new version of the website on a small percentage of users before fully releasing it. This is called a canary deployment. You will configure Nginx to route 10% of the traffic to the new version and 90% to the stable version.

🎯 Goal: Build an Nginx configuration that splits incoming user traffic so that 10% goes to the canary backend server and 90% goes to the stable backend server. This helps test new changes safely.

📋 What You'll Learn

Create an upstream block with two backend servers named stable_backend and canary_backend

Add a variable canary_weight set to 10 representing 10% traffic for canary

Use the split_clients directive to split traffic based on $remote_addr into canary and stable

Configure the server block to proxy requests to the correct backend based on the split

Print the final Nginx configuration that routes traffic accordingly

💡 Why This Matters

🌍 Real World

Canary deployments are used in real companies to reduce risk when releasing new software versions by testing changes on a small user group first.

💼 Career

DevOps engineers often configure load balancers like Nginx to manage traffic routing for canary deployments, improving software reliability and user experience.

1

Create upstream backend servers

Create an upstream block named backend with two servers: stable_backend at 192.168.1.10:80 and canary_backend at 192.168.1.20:80.

Nginx

# Your code here

Need a hint?

Use the upstream directive to group backend servers.

2

Add canary traffic weight variable

Add a variable canary_weight and set it to 10 to represent 10% traffic for the canary deployment.

Nginx

upstream backend {
    server 192.168.1.10:80;  # stable_backend
    server 192.168.1.20:80;  # canary_backend
}

# Your code here

Need a hint?

Use the map directive to create a variable based on $remote_addr.

3

Split traffic using split_clients

Use the split_clients directive with $remote_addr to split traffic into canary and stable groups. Assign canary to 10% and stable to 90%.

Nginx

upstream backend {
    server 192.168.1.10:80;  # stable_backend
    server 192.168.1.20:80;  # canary_backend
}

map $remote_addr $canary_weight {
    default 10;
}

# Your code here

Need a hint?

The split_clients directive helps distribute traffic by percentage.

4

Configure server block to route traffic

Create a server block listening on port 80. Use an if statement to proxy requests to canary_backend if $deployment_group is canary, otherwise proxy to stable_backend. Use proxy_pass with http://backend and the correct server.

Nginx

upstream backend {
    server 192.168.1.10:80;  # stable_backend
    server 192.168.1.20:80;  # canary_backend
}

map $remote_addr $canary_weight {
    default 10;
}

split_clients "$remote_addr$canary_weight" $deployment_group {
    10% canary;
    * stable;
}

# Your code here

Need a hint?

Use if inside location to route traffic based on $deployment_group.