Flo&Instreet E-Commerce AWS EKS(Elastic Kubernetes Service) Case Study -(ENG)
Flo and Instreet, which are the most important brands among the Ziylan group, develop their e-commerce infrastructure on Magento2. Ziylan Group leads the retailing field, and the organization wants to maintain the same performance in e-commerce and mobile channels. Our most important key words in this project are; scalability, high availability and zero down time deployment. Especially zero downtime deployment is a huge challenge for Magento2. We completed the project in 3.5 weeks with the support and agile approach of Gökhan Uygun, The Director of IT at Ziylan Group, and Ozan Akçora from RG Tech, who lead the software developers. We would like to thank all our colleagues who contributed to the project.
- We make the existing structure more flexible by making it functional on container.
- We used AWS EKS service for the container cluster.
- We used CloudFormation for EKS worker nodes scalability operations.
- We used helm on Kubernetes deployment processes and we used S3 service for helm repository. We also used ECR Service for the container images that we build.
- We used Ansible and Jenkins tools for orchestration on CI/CD processes with the perspective of pipeline as a code.
- We used Prometheus, Alert Manager, Grafana for the metric monitoring. And we used CloudWatch, SNS, Lambda for the EKS and worker node metrics.
- We used AWS ElasticSearch Service, Fluentd, Graylog tools for the log infrastructure.
- We used Kubernetes hpa for scaling of the application, and for scaling of the worker node we used Cluster Autoscaler and AWS AutoScaling Group.
- We gained a scalable, flexible and high available infrastructure.
- We gained the opportunity of zero down time deployment at any moment.
- We gained the ability of detecting a problem before it occurs with the help of centralize log management and metric monitoring structure.
- We gained the cost optimization by using spot instance on worker nodes scaling.