Launching Services in
Amazon Web Services
James Armes, Cloud Software Engineer, Acquia
– Automated code deployments.
– Agent is available for Linux or
– Can be used for instances in AWS or
– Repeatable deployments.
– Rolling updates with rollback.
– State tracker & task coordinator.
– Promotes separation of control-flow
and business logic.
– Tasks can be run in parallel or
– Jobs can be picked by in instances
in AWS or on-premises.
– Create/manage a collection of
– Create reusable templates.
– Avoids collisions.
Cloud Database Service
– Allow internal users to create
– Provide various service levels.
– Abstracts hardware and
– Maintenance is largely automated.
– Resources are isolated in separate
Cloud Database Service (cont.)
– Custom API and worker AMIs.
– Bugsnag for reporting application
– Logs sent to Sumo Logic.
– StatsD/Graphite for performance
– Deployments managed through
Cloud Database Service: API
– HTTP API built with grape.
– Public and internal endpoints with
– HMAC authentication.
Cloud Database Service: Worker
– Picks up workflows and activities
– Built with aws-flow gem.
– Workflows utilize atomic activities.
– Activities are idempotent.
The cloud is the Internet. Rather than a local or collocated server, your servers are located “somewhere” on the Internet. You can specify a “region” for your servers. The exact location of the hardware is unknown to you. Pay for only the resources you use. On-demand capacity. Prepay reservations available. Elastic capacity. Scale your resources up and down based on usage requirements.
Code Deploy: Integrates with ASGs Repeatable deployments Deployments are uploaded to S3
Simple Workflow Service
CloudFormation: Template JSON format Can be parameterized Collisions Resources are given a unique name