Using Docker Images

Deploy an application using a container image stored in your Drycc Container Registry.

Drycc supports deploying applications via an existing Docker Image. This is useful for integrating Drycc into Docker-based CI/CD pipelines.

Prepare an Application

Start by cloning an example application:

$ git clone
$ cd example-dockerfile-http

Next use your local docker client to build the image and push it to DockerHub.

$ docker build -t <username>/example-dockerfile-http .
$ docker push <username>/example-dockerfile-http

Docker Image Requirements

In order to deploy Docker images, they must conform to the following requirements:

  • The Dockerfile must use the EXPOSE directive to expose exactly one port.
  • That port must be listening for an HTTP connection.
  • The Dockerfile must use the CMD directive to define the default process that will run within the container.
  • The Docker image must contain bash to run processes.

!!! note Note that if you are using a private registry of any kind (gcr or other) the application environment must include a $PORT config variable that matches the EXPOSE’d port, example: drycc config:set PORT=5000. See Configuring Registry for more info.

Create an Application

Use drycc create to create an application on the controller.

$ mkdir -p /tmp/example-dockerfile-http && cd /tmp/example-dockerfile-http
$ drycc create example-dockerfile-http --no-remote
Creating application... done, created example-dockerfile-http

!!! note For all commands except for drycc create, the drycc client uses the name of the current directory as the app name if you don’t specify it explicitly with --app.

Deploy the Application

Use drycc pull to deploy your application from DockerHub or a public registry.

$ drycc pull <username>/example-dockerfile-http:latest
Creating build...  done, v2

$ curl -s
Powered by Drycc

Because you are deploying a Docker image, the web process type is automatically scaled to 1 on first deploy.

Use drycc scale web=3 to increase web processes to 3, for example. Scaling a process type directly changes the number of Containers running that process.

Private Registry

To deploy Docker images from a private registry or from a private repository, use drycc registry to attach credentials to your application. These credentials are the same as you’d use when running docker login at your private registry.

To deploy private Docker images, take the following steps:

  • Gather the username and password for the registry, such as a Robot Account or a Long Lived Token
  • Run drycc registry:set username=<the-user> password=<secret> -a <application-name>
  • Now perform drycc pull as normal, against an image in the private registry

When using a Long Lived Token, the JSON blob will have to be compacted first using a tool like jq and then used in the password field in drycc registry:set. For the username, use _json_key. For example:

drycc registry:set username=_json_key password="$(cat google_cloud_cred.json | jq -c .)"

When using a private registry the docker images are no longer pulled into the Drycc Internal Registry via the Drycc Workflow Controller but rather is managed by Kubernetes. This will increase security and overall speed, however the application port information can no longer be discovered. Instead the application port information can be set via drycc config:set PORT=80 prior to setting the registry information.

!!! note Currently and ECR in short lived auth token mode are not supported.