220 lines (157 sloc) 9.24 KB

Getting Started

This document shows how to deploy a "Hello World" application with Draft. If you haven't done so already, be sure you have Draft installed properly. This Quickstart Guide is the perfect resource if you still need to install Draft.

Application Setup

There are multiple example applications included within the examples directory. For this walkthrough, we'll be using the python example application which uses Flask to provide a very simple Hello World webserver.

$ cd examples/example-python

Draft Create

We need some "scaffolding" to deploy our application into a Kubernetes cluster. Draft can create a Helm chart, a Dockerfile, and a draft.toml with draft create:

$ draft create
--> Draft detected Python (96.739130%)
--> Ready to sail
$ ls -a
.dockerignore     .draftignore            draft.toml
.draft-tasks.toml Dockerfile        charts/           requirements.txt

The charts/ and Dockerfile assets created by Draft default to a basic Python configuration. This Dockerfile harnesses the python:onbuild image, which will install the dependencies in requirements.txt and copy the current directory into /usr/src/app. To align with the internalPort service value in charts/python/values.yaml, this Dockerfile exposes port 8080 from the container.

The draft.toml file contains basic configuration details about the application like the name, the repository, which namespace it will be deployed to, and whether to deploy the application automatically when local files change.

$ cat draft.toml
    name = "example-python"
    namespace = "default"
    wait = true
    watch = false
    watch-delay = 2
    auto-connect = false
    dockerfile = ""
    chart = ""

See for more information and available configuration on the draft.toml file.

A .draftignore file is created for elements we want to exclude tracking on draft up when watching for changes. The syntax is identical to helm's .helmignore file.

$ cat .draftignore

A .dockerignore file is created to ensure the docker context ignores files and directories that are not necessary.

$ cat .dockerignore

A .draft-tasks.toml file is also created. This file allows you to configure tasks to be run before draft up (pre-up tasks), after draft up (post-up tasks), or after draft delete (post-delete tasks). This file is empty by default. See for more information and available configuration on the .draft-tasks.toml file.

Draft Up

Now we're ready to deploy this application to a Kubernetes cluster. Draft handles these tasks with one draft up command:

  • reads configuration from draft.toml
  • compresses the charts/ directory and the application directory as two separate tarballs
  • builds the image using docker
  • instructs docker to push the image to the registry specified in draft.toml (or in draft config get registry, if set)
  • instructs helm to install the chart, referencing the image just built
$ draft up
Draft Up Started: 'example-python': 01BSY5R8J45QHG9D3B17PAXMGN
example-python: Building Docker Image: SUCCESS ⚓  (52.1337s)
example-python: Releasing Application: SUCCESS ⚓  (0.5309s)
Inspect the logs with `draft logs 01BSY5R8J45QHG9D3B17PAXMGN`

NOTE: You might see a WARNING: no registry has been set message if no container registry has been configured in draft. You can set a container registry using the draft config set registry command. If you'd prefer to silence this warning instead, you can run draft config set disable-push-warning 1.

To ensure your application deployed as expected, run kubectl get pods and take a look at the output.

$ kubectl get pods
NAME                                     READY     STATUS    RESTARTS   AGE
example-python-python-6755c4944d-zbgvj   1/1       Running   0          5s

NOTE: If you're using Minikube and your STATUS shows an error such as ErrImagePull or ImagePullBackOff, make sure you've configured Draft to build images directly using Minikube's Docker daemon. You can do so by running eval $(minikube docker-env).

INFO: For more information on installing and configuring Minikube for use with Draft, check out the Minikube installation guide here.

Interact with the Deployed Application

Now that the application has been deployed, we can connect to it using draft connect.

The draft connect command is used to interact with the application deployed on your cluster. It works by creating proxy connections to the ports exposed by the containers in your pod. It also streams the logs from all containers.

$ draft connect
Connect to python:8080 on localhost:54794
[python]:  * Environment: production
[python]:    WARNING: Do not use the development server in a production environment.
[python]:    Use a production WSGI server instead.
[python]:  * Debug mode: off
[python]:  * Running on (Press CTRL+C to quit)

NOTE: The WARNING: Do not use the development server in a production environment message is coming from Flask. The message is in regard to Flask's built-in web server and can safely be ignored for our test purposes here.

In this example, you can see that draft connect has proxied port 8080 from our container to port 54794 on localhost. We can now open a browser window or another terminal window and connect to our application using the address and port displayed from draft connect's output.

$ curl localhost:54794
Hello, World!

IMPORTANT: Your local port will likely be different than the one seen here.

NOTE: If localhost does not resolve on your system, try curl<PORT> instead.

Once you're done checking the application out, you can cancel out of the draft connect session using CTRL+C.

NOTE: You can use the flag draft up --auto-connect in order to have the application automatically connect once the deployment is done.

INFO: You can also customize the local ports for the draft connect command by using the -p flag or through the override-ports field in draft.toml. More information on this can be found in

Update the Application

Now, let's change the output in to output "Hello, Draft!" instead:

$ cat <<EOF >
from flask import Flask

app = Flask(__name__)

def hello():
    return "Hello, Draft!\n"

if __name__ == "__main__":'', port=8080)

Draft Up(grade)

When we call draft up again, Draft determines that the Helm release already exists and will perform a helm upgrade rather than attempting another helm install:

$ draft up
Draft Up Started: 'example-python': 01CEQ5H21BWSR5M8HTJ5BVXPYW
example-python: Building Docker Image: SUCCESS ⚓  (1.0010s)
example-python: Releasing Application: SUCCESS ⚓  (2.1236s)
Inspect the logs with `draft logs 01CEQ5H21BWSR5M8HTJ5BVXPYW`

We should notice a significantly faster build time here. This is because Docker is caching unchanged layers and only compiling layers that need to be rebuilt in the background.

Great Success!

We can run draft connect again to set up a proxy to our application:

$ draft connect
Connect to python:8080 on localhost:54961
[python]:  * Environment: production
[python]:    WARNING: Do not use the development server in a production environment.
[python]:    Use a production WSGI server instead.
[python]:  * Debug mode: off
[python]:  * Running on (Press CTRL+C to quit)

Once we have the address and port, we can connect again using curl in a new terminal window or by browsing to the host and port in a browser window:

$ curl localhost:54961
Hello, Draft!

We can see the application updated successfully!

Draft Delete

If you're done testing this application, you can terminate and remove it from your Kubernetes cluster. To do so, run draft delete:

$ draft delete
app 'example-python' deleted

If you run kubectl get pods shortly after, you should see your application STATUS is Terminating:

$ kubectl get pods
NAME                                     READY     STATUS        RESTARTS   AGE
example-python-python-688fcf849f-8ddh7   1/1       Terminating   0          5m

Once the termination completes, a kubectl get pods will show that the application no longer exists in your Kubernetes cluster:

$ kubectl get pods
No resources found.

IMPORTANT NOTE: The draft delete command should be run with extreme care and caution as it performs a termination and removal of the application from your Kubernetes cluster.

INFO: The draft delete command does not any image(s) created for the deployment within your Docker registry.