Achieved a 25% Cost Reduction with Just 0.1% Downtime for an E-Commerce Company

Solutions

• Applications are deployed across distinct environments (development, staging, pre-production, production) with workloads orchestrated on Amazon ECS. AWS CloudFormation, an Infrastructure as Code (IaC) tool, is used to provision and manage these environments. 

• Bitbucket and Jenkins handle version control and continuous integration. A Jenkins pipeline automates app building, Dockerization, and storage on Amazon S3, while AWS CodePipeline manages Amazon ECR changes and Amazon ECS deployments through CodeDeploy. 

• Amazon CloudWatch, Prometheus, and Glowroot are used for APM and infrastructure monitoring. They track metrics from Amazon ECS Cluster instances, Amazon EC2 instances, and Amazon ECS tasks. This includes monitoring memory, CPU, and disk utilization with an integrated email alerting system. 

• Amazon ECS Service Auto Scaling is implemented to dynamically adjust the desired task count based on changing demand, with scaling policies triggering scaling events. Auto Scaling Groups manage the load for Amazon EC2 instances. 

• A load balancer distributes traffic among Amazon ECS tasks. An internet-facing Application Load Balancer (ALB) handles web layer exposure, while internal ALBs balance traffic for the internal application servers. 

• Amazon RDS with Multi-AZ deployment ensures high availability and automated failover, with scheduled backups, replication, and snapshots for data durability and recovery. 

• Build artifacts are stored in Amazon S3, with Amazon ECR used as a Docker image registry.