Benefits of using Docker Containers

containers.md

• Package format that not only contains our code, but also the entire definition of the OS that runs the code which means you get guaranteed consistency in your deployments
• Easy deployments
  • Just tell docker to run an image from our build artifact registry
• Simple build promotion
• Easy scaling
• Run console workers that will automatically be restarted based on a restart policy
• Cheaper because we can pack more services onto a box and utilize resources more efficiently while still providing isolation and cpu / memory constraints
• Can scale each worker independently
• Docker registry for private docker build artifact storage
• All kinds of apps already have docker images setup and we would be able to deploy those things to our environments easily.
  • Jenkins, Elasticsearch, Redis, etc
• All kinds of base images to derive from for things like java8, iis, etc
• Can run both linux and windows containers in the same docker swarm cluster
• Can run full .net framework on windows containers and eventually nano for .net core or even linux for .net core apps
• Can direct all log output into a central log store
• Can collect base level metrics like cpu and memory usage in a central location for each monitoring
• Can run docker containers locally on windows and osx and attach debuggers to apps running inside of the docker containers
• Manage everything we do in the same way while not locking ourselves into any cloud provider
• Docker compose can be used to configure a set of containers that work together
  • We would have 2 web apps, redis, elasticsearch and multiple console workers
• Ability to isolate and run each worker process as a seperate container with seperate monitoring and logging and scaling
• There are several free container orchestration projects to choose from so really we are just locking ourselves into using containers and not much other lock in
• Both AWS and Azure have container services that make managing the cluster and orchestration simple