New container types are being put into operation in cooperation with the administrators of root360. The process is explained below. It lists what information needs to be provided for root360 to set up a new container type.

Process

• Request a new container type via a customer portal https://support.root360.cloud and provide required information (see questionnaire below)

• Discuss questions and technical details with root360

• When the requirements are clear, root360 will integrate a set of instances of the new container type into your existing Elastic Container Service (ECS) cluster

• Test the instance(s)/service(s)

• Go-live and possibly scale the container type

New Container questionnaire

• Short, unique container name (e.g. frontend, admin):

• Short description of the container’s purpose:

• Resource requirements for the container instances in CPU-Units (1000 = 1 vCPU), RAM in MiB:

• How many container instances? Should we configure Scale-out/Autoscaling?

• If container autoscaling: minimum+maximum number of containers?

• On which port and with which protocol (http, https or tcp) will the container image listen on?

• If accessibility from the public internet: through which domain(s) and/or URL-Paths should the container type be accessible:

• Name a container healthcheck path. We check this path for a http(s) return code (usually 200) to continually validate the container state:

• Is a mount-point needed as a share?

• What is the volume of data and the requirements for the performance of the data storage?

• List required environment variables, e.g. database credentials, endpoints of other containers
  Do not send secrets through the customer service portal or through any other plain text communication channel. Please use https://otp.skaylink.com/ to ensure a secure, encrypted one-time transmission.

• Is access to other components needed (e.g. databases or S3 buckets)?

• Which kind of deployment is used? (blue/green or replacement)

• Logging: we use Cloudwatch by default. Do you want to use another logdriver? (Fargate: only Cloudwatch supported) (For detailed description of logging for docker see ​​​​​​​​Understanding container logging configuration.)

Related tutorials

Related components