Deployment Overview

There are many ways you can run Bytewax, in this section we will highlight the main ways we recommend deploying Bytewax.

The Bytewax Platform

We built the Bytewax Platform to make deploying, managing and operating dataflows easier. You can find out more about the platform features on the Bytewax website and the platform documentation.

Container Based Deployment

Docker containers provide a good way to wrap up dependencies and run a containerized version of Bytewax across many different services and orchestration tools. From docker compose to Kubernetes. The container documentation provides an overview of how to leverage Bytewax in containers.

Running Locally with waxctl

We recommend using waxctl locally to test running your dataflow with multiple processes, you can find more about waxctl on the website.

Test it out locally after installing:

$ waxctl run mydataflow.py -p 2

You can find more information in <project:#ref-waxctl>.

## Running in Kubernetes

Kubernetes is the most common orchestration platform for running containers. The recommended path to using Bytewax with Kubernetes is to use the [platform](https://platform.bytewax.io/) but you can use the publicly available [helm chart](#xref-helm) to run Bytewax on a Kubernetes cluster.

## Running on Cloud Instances

You can deploy a dataflow directly to a cloud instance on [AWS](#xref-waxctlaws) and [GCP](#xref-waxctlgcp) with `waxctl` doing the heavy lifting of setting up the machine and running the dataflow.