Amazon ECS Workshop > Blue/Green Deployments > Rollback failed deployment

Rollback failed deployment

Introduce a breaking change

Here we will trigger an automated rollback by returning a 500 response code. The code deployment group has the CloudWatch alarm configured to detect and rollback a deployment with 5xx errors.

Edit the nginx.conf to return a 500 error instead of default index.html

Modify the / location directive to look like the following:

listen  80;
root    /usr/share/nginx/html;
include /etc/nginx/mime.types;

location / {
    return 500;
}

Expand to see the complete nginx.conf

# Run as a less privileged user for security reasons.
user nginx;

# Number of worker_threads to run;
# "auto" sets it to the #CPU_cores available in the system, and
# offers the best performance.
worker_processes    auto;

events { worker_connections 1024; }

http {
    server {
        # Hide nginx version information.
        server_tokens off;

        listen  80;
        root    /usr/share/nginx/html;
        include /etc/nginx/mime.types;

        location / {
            return 500;
        }

        gzip            on;
        gzip_vary       on;
        gzip_http_version  1.0;
        gzip_comp_level 5;
        gzip_types
                        application/atom+xml
                        application/javascript
                        application/json
                        application/rss+xml
                        application/vnd.ms-fontobject
                        application/x-font-ttf
                        application/x-web-app-manifest+json
                        application/xhtml+xml
                        application/xml
                        font/opentype
                        image/svg+xml
                        image/x-icon
                        text/css
                        text/plain
                        text/x-component;
        gzip_proxied    no-cache no-store private expired auth;
        gzip_min_length 256;
        gunzip          on;
    }
}

Push the code to the CodeCommit repository

cd ~/environment/nginx-example
git add .
git commit -m "Returning 500 error"
git push

The code push will trigger the CodePipeline

Rollback-Deployment

The CodeDeploy initiates the deployment as before
Once the step 3 traffic shifting is started, we open the service in your browser on the Load Balancer port 80

Here is the command to get the url:

echo "http://$ALB_DNS"

You will notice the page returning 500 Internal Server error
We have configured CloudWatch alarms which will look for 5XX errors from the ELB target groups

Expand to see the code

The CodeDeploy is configured to initiate automatic rollback if the CloudWatch alarms are In Alarm state
The CodeDeploy will stop the deployment
CodeDeploy will now re-route the production traffic to original task set
Finally, CodeDeploy will terminate the replacement task set
Refresh the service on the browser and the original green background deployment will be visible

That’s it. You have successfully completed a blue/green deployment and done automatic rollback of a failed deployment. Let’s review the configurations files now.