Engine Yard

Git push deployment on AWS

Customers with little DevOps capacity can ship code without managing servers, because Engine Yard handles the infrastructure beneath their web applications on Amazon Web Services. Its core workflow is a Git push: a developer commits and pushes code to a remote repository, and the platform analyzes it, builds containers and runs them on Kubernetes. The company also deploys the utilities that handle background jobs and cron tasks.

The company has a long background in Ruby and Rails applications. Other stacks it lists include Node.js, PHP, Java and Python, together with database support and use cases such as cloud application hosting. It positions itself as an alternative to Heroku, OpenShift and Elastic Beanstalk, and it publishes comparison pages for each.

Scaling and isolation

Each customer receives a dedicated AWS account and a set of virtual networks to avoid the noisy neighbor problem, in which one tenant's load slows another's. The Kubernetes cluster is private and fully configured, and the company says compliance requirements are built into the platform at no extra charge. Scaling reacts to demand, with pods that scale horizontally on processor, memory and custom measures, and pod sizes set in 1GB increments. Stacks are pre-optimized for each application type. The platform also uses machine learning to tune infrastructure from monitored performance, which the company says reduces costs.

Updates, backups and monitoring

Engine Yard handles stack and security updates, and it plans replication, backups and recovery. Rolling deployments are designed to release code without downtime or a drop in performance. Logs from every application are gathered in one place using the ELK stack (Elasticsearch, Logstash and Kibana), and each application connects to a Grafana instance that shows metrics such as memory and CPU use. The platform configures security groups and firewalls to limit exposure.