README.md

Akibox Tutorial

This is a tutorial project to help you get to know Akita. It contains a FastAPI server implementing a toy Dropbox-like file server. You can use Akita to generate a spec for its API and track API performance.

The Akita Client works by capturing traffic to your service -- Akibox, in this case. This tutorial showcases a few different ways to run Akibox and the Akita Client.

Before you get started, you'll need to create a new project in the Akita app. Follow these instructions. Name your project akibox.

Run Akibox on the Host

Follow these instructions to run Akibox directly on your host and to capture host traffic with the Akita Client.

Create a python virtual environment and install dependencies

python3 -m venv venv
source ./venv/bin/activate
pip install -r requirements.txt

Start the Akibox service

uvicorn main:app --reload

Start the Akita Client

If you haven't already, install the Akita Client.

Next, open another terminal window and start the client.

akita apidump --service akibox --filter "port 8000"

Make some requests!

Use the test.sh script to make some requests against your service.

./test.sh

Run Akibox with Docker

Optionally, you can build Akibox into a Docker container. In this case, there are a few ways to run the Akita Client:

1. In a separate Docker container attached to the Akibox container's network.
2. In the same Docker container, as a wrapper that invokes Akibox.
3. In the same Docker container, as a background process managed by s6.

Build Akibox in a Separate Docker Container

docker build -t akibox-tutorial .

Run Akibox

docker run -e PORT=8000 -p 8000:8000 --name akibox-tutorial akibox-tutorial

Start the Akita Client

Open another terminal window and start the client in a Docker container.

docker run --rm --network container:akibox-tutorial \
  -e AKITA_API_KEY_ID=your-api-key \
  -e AKITA_API_KEY_SECRET=your-secret \
  akitasoftware/cli:latest apidump --service akibox --filter "port 8000"

Build Akibox and the Akita Client in the Same Container

There are two ways to run the Akita Client as a background process in a Docker container. To test out Akita, you can wrap your service with the Akita Client. The Dockerfile.withcli file shows how this works.

There are two drawbacks to this approach, however.

1. If the Akita Client encounters an error, it will stop your service as well as itself.
2. The Akita Client does not pass signals it receives to your service. It will stop if it receives SIGINT or SIGTERM.

As a more robust alternative, you can use s6-overlay to run the Akita Client as a background process. The Dockerfile.withs6 file shows how to do this.

Once you choose an approach, build the akita-tutorial image.

docker build -t akibox-tutorial -f Dockerfile.withcli .

Run Akibox and the Akita Client

docker run -e PORT=8000 -p 8000:8000 --name akibox-tutorial \
  -e AKITA_API_KEY_ID=your-api-key \
  -e AKITA_API_KEY_SECRET=your-secret \
  akibox-tutorial

Make some requests!

Once you've started Akibox and the Akita Client, use the test.sh script to make some requests against your service.

./test.sh

Run Akibox with Docker Compose

You can also use Docker Compose to start Akibox. In this case, the Docker Compose file can also start the Akita Client in separate container attached to the same Docker network.

docker-compose up

Make some requests!

Use the test.sh script to make some requests against your service.

./test.sh

Integration Tests

Integration tests are defined in test_main.py. To run them:

pytest