LCP

Blue-Green Deployment using Azure for B2B online platform

undefined
Overview

Blue-Green deployment for brand and reseller B2B collaborative platform serving heavy and consistent traffic

At A Glance

industry
Industry
Retail & E-commerce
region
Region
USA
duration
Duration
2 Weeks

Technical Stack

postgre-sql
azure-vm
azure-traffic-manager
azure-resource-group
azure-load-balancer
azure-cloud
elasticsearch

Client Profile

Based out of the USA, the client is a PaaS provider with a platform for B2B online portal, software platform, transaction engine, and secure marketplace.

Challenge

  1. The all-new platform had been introduced by the client, which required a smooth deployment without any downtime for the users. This was an issue for the client.
  2. The client wanted to test the platform with a few users before deploying it to the mass without affecting the user experience.
  3. The client wanted to perform A/B testing for both platforms to check whose performance was better.
  4. There was no bandwidth available with the client for database, resource, and performance management, along with system administration for the deployment.
Azure's blue-green deployment process diagram for B2B platform, demonstrating parallel environments and controlled updates for reliability

Solution

  1. Seaflux created two production environments, one for the current version (blue) and one for the new (green), on the Azure cloud.
  2. Deployment of the new version is done using the Blue-Green Deployment method, where we used the PostgreSQL flexible server as the database for better control and customizations.
  3. The Blue-Green Deployment method helped to test the new platform on a few users before deploying the whole build. To do this, we used the Traffic Manager and Load Balancer in Azure to manage & control the traffic. The traffic was limited to 10% initially to the green deployment in order to test the new platform. This allowed simple rollbacks in case there remains any bugs in the new build.
  4. Once the green deployment was tested, and there were no rollbacks, half the traffic was gradually shifted from the blue to the green deployment for A/B testing.
  5. After proper calculation and analysis of the performance of users, the whole traffic was diverted to the best-selected platform to provide better usability. 

Key Benefits

  1. Successful and rapid release of the new platform with 100% uptime.
  2. The method allowed simple rollbacks and built-in disaster recovery without impacting the users.
  3. Using PostgreSQL Flexi Server allows scalability, automatic backups, security, and monitoring to the client.

Develop your next idea with usGet in touch