headphonejames / armada

A multi-cluster batch queuing system for high-throughput workloads on Kubernetes.

Armada

Armada is a multi-Kubernetes cluster batch job scheduler.

Armada is designed to address the following issues:

1. A single Kubernetes cluster can not be scaled indefinitely, and managing very large Kubernetes clusters is challenging. Hence, Armada is a multi-cluster scheduler built on top of several Kubernetes clusters.
2. Acheiving very high throughput using the in-cluster storage backend, etcd, is challenging. Hence, queueing and scheduling is performed partly out-of-cluster using a specialized storage layer.

Armada is designed primarily for machine learning, AI, and data analytics workloads, and to:

• Manage compute clusters composed of tens of thousands of nodes in total.
• Schedule a thousand or more pods per second, on average.
• Enqueue tens of thousands of jobs over a few seconds.
• Divide resources fairly between users.
• Provide visibility for users and admins.
• Ensure near-constant uptime.

Armada is a CNCF Sandbox project used in production at G-Research.

For an overview of Armada, see this video.

Documentation

For an overview of the architecture and design of Armada, and instructions for submitting jobs, see:

• System overview
• User guide

For instructions of how to setup and develop Armada, see:

• Development guide

For API reference, see:

• API Documentation

We expect readers of the documentation to have a basic understanding of Docker and Kubernetes; see, e.g., the following links:

• Docker overiew
• Kubernetes overview

Armada follows the CNCF Code of Conduct