Deploying Open Source Software on Azure: New Guides for AKS and VMs

“Tech companies born with an open-source mentality get it. It’s our ability to work together that makes our dreams believable, and ultimately achievable; we must learn to build on the ideas of others.”—Satya Nadella, CEO, Microsoft.

For well over a decade Azure has invested in integrating open source deeply into its services and ecosystems, and this continues to accelerate.

To accelerate this integration further and respond to user needs, the customer experience team has developed end-to-end guides for deploying open source software on Azure Kubernetes Service (AKS) and Azure virtual machines (VMs). The guides span an initial selection of projects: PostgreSQL, Valkey, Karpenter-based NAP, Flyte, KAITO, and Inspektor Gadget to empower Stateful, AI, and other common workloads.

Why did we make these guides?

Deploying and managing open source software on a cloud platform can come with its challenges. To address this, these guides are designed to help you get some of the most critical and heavily adopted open source projects onto Azure using good practice guidelines as quickly as possible.

These guides provide clear, step-by-step instructions, troubleshooting tips, and optimization strategies. To make things even easier, some of these best practice guides can be automatically deployed from Portal, so that you can quickly get started or improve your learning experience.

Best Practice Deployments across Workloads

The guides cover established projects such as, PostgreSQL, Valkey (Redis fork), and WordPress and highlight new and exciting projects like KAITO and Flyte to enable open source AI application development. We also provide guidance on implementing critical capabilities like Karpenter-based AKS Node autoprovisioning scaling capability and Inspektor Gadget to gather real-time kernel-level insights using eBPF.

Below is a brief overview of each guide available today. In addition to basic installation steps, the guides detail best practice recommendations on networking and security. Additionally, if deployed on Kubernetes, you will also find recommendations on cluster architecture. These guides walk you through the entire process, from initial setup to advanced configurations, ensuring that even those with limited cloud experience can successfully deploy their applications with best practices enabled by default.

Stateful Workloads

PostgreSQL is a staple of modern cloud native application development. Run PostgreSQL on Kubernetes with the same high availability and multi-region redundancy offered by managed or VM deployments. Deploy a highly available PostgreSQL database on AKS.

A high-performance key value datastore like Valkey supports a variety of workloads such as caching, message queue, or a primary database. Deploy a Valkey Cluster on AKS.

AI Workloads

The AI toolchain operator (KAITO) automatically provisions the necessary GPU nodes and sets up the associated inference server as an endpoint server to your AI models on your AKS clusters. Deploy an AI model on Azure Kubernetes Service (AKS) with the AI toolchain operator (preview).

Quickly build and test complex workflows to train small models, scale GPU-accelerated deep learning, or improve statistical summarization with Flyte. Deploy data and machine learning pipelines with Flyte on Azure Kubernetes Service (AKS).

General Workloads

Deploy a website or blog built on the LEMP stack with WordPress in one click. Deploy a LEMP stack using WordPress on a VM.

Newly built on the open source Kubernetes autoscaler, Karpenter, Node autoprovisioning (NAP) for AKS optimizes compute resources for better application availability and efficiency. Node autoprovisioning (preview) – Azure Kubernetes Service | Microsoft Learn.

Efficiently debug and inspect infrastructure in real time using Inspektor Gadget which is a tool and framework for data collection on Kubernetes clusters and Linux hosts, using eBPF. Capture real-time system insights from an AKS cluster – Azure | Microsoft Learn.

What’s Next?

Azure is committed to helping our customers build and run workloads on the opens source solutions that they depend on. We at Azure will continue to expand our library of best practice deployment guides to ensure that developers can deploy OSS across Azure Kubernetes Service (AKS) and Azure Virtual Machines (VMs). And there is more to come!

You can expect to see more from us across a variety of open source projects like Apache Kafka, Apache Airflow, and more.