Quickstart

Get started with Chamber in under 10 minutes

This guide walks you through setting up Chamber for your organization, from requesting access to submitting your first workload.

Prerequisites

• A Kubernetes cluster with GPU nodes
• kubectl access to your cluster
• helm v3 installed

Don’t have a GPU Kubernetes cluster yet? Use our Terraform modules to deploy a GPU-ready cluster on AWS or GCP in ~15 minutes.

Step 1: Create Your Organization

1. Request access

   Contact us at [e-mail] to request access to Chamber.

2. Create organization

   After signing in, you’ll be prompted to create an organization. This is your top-level tenant in Chamber.

3. Invite team members

   Navigate to Settings > Members to invite colleagues to your organization.

Step 2: Connect Your Cluster

Install the Chamber agent in your Kubernetes cluster to sync workload state with the control plane.

# Get your cluster token from the Chamber dashboard
# Navigate to Settings > Security > API Tokens -> New Token

# Install the agent
helm install chamber-agent oci://public.ecr.aws/chamber/chamber-agent-chart \
  --version 0.8.96 \
  --namespace chamber-system \
  --create-namespace \
  --set saas.url="wss://controlplane-api.usechamber.io/agent" \
  --set saas.token="<YOUR_TOKEN>" \
  --set saas.clusterId="your-gpu-cluster"

Verify the agent is running:

kubectl get pods -n chamber-system -l app.kubernetes.io/name=chamber-agent

Step 3: Create Your First Team

Teams represent teams or projects in your organization. They form a hierarchy for organizing capacity allocation.

1. Navigate to Teams

   In the Chamber dashboard, click Teams in the sidebar.

2. Create root team

   Click Create Team and enter:

   • Name: e.g., “ML Platform”
   • Description: Brief description of the team/project
3. Add projects (optional)

   Create sub-teams for different teams or projects under your root team.

Step 4: Allocate Capacity

Reserve GPU capacity for your team from your connected cluster.

1. View capacity pools

   Navigate to Capacity Pools to see your connected clusters.

2. Create reservation

   Click on a pool, then Create Reservation:

   • Select the team to receive capacity
   • Specify the number of GPUs to reserve

Start with a small reservation and adjust based on utilization metrics. You can always increase or redistribute capacity later.

Step 5: Submit a Workload

With capacity reserved, your team can now submit GPU workloads.

Example Workload

apiVersion: batch/v1
kind: Job
metadata:
  name: training-workload
  labels:
    chamber.io/team: ml-platform       # Links workload to team
    chamber.io/workload-class: reserved # Use reserved capacity
spec:
  template:
    spec:
      containers:
      - name: trainer
        image: your-training-image:latest
        resources:
          limits:
            nvidia.com/gpu: 4
      restartPolicy: Never

Workloads without the chamber.io/workload-class label default to elastic and will use idle capacity.

Next Steps

Understand Teams

Learn how to structure your organization hierarchy

Workload Classes

Understand reserved vs elastic workloads

Dashboard Guide

Monitor utilization and workload status