Stackable Docs Hub

Stackable

Stackable

GitOps in action with ArgoCD

Demo · argo-cd-git-ops

GitOps in action with ArgoCD

Deploy and operate the Stackable Data Platform declaratively, with every change living in Git, reviewed like code, and continuously reconciled onto your cluster by ArgoCD.

HOW TO START

With a Kubernetes cluster, run a single command:

$stackablectl demo install argo-cd-git-ops

Key Features

What this demo shows

Declarative platform

In this demo, Airflow from our stack is described as a manifest in Git, with no click-ops.

Auditable history

Every change is a commit, showing who changed what and when, with instant rollback.

Self-healing

Manual edits to the cluster are reverted to match the repository.

Continuous reconciliation

ArgoCD watches Git and automatically syncs any drift back to the desired state.

Environment parity

Promote the exact same manifests from dev to staging to production.

App-of-apps

A single root application bootstraps operators, data products and dashboards.

Demo Components

The stack under the hood

ArgoCD

GitOps engine

Continuously reconciles Kubernetes state against the Git repository, with a web UI showing sync and health of every resource.

Apache Airflow

Orchestration

Scheduled data workflows, deployed the same declarative way as the rest of the platform.

MinIO

S3 storage

S3-compatible object store backing the demo's data products.

Stackable Operators

Platform

All installed and configured through the same GitOps flow.

Demo Workflow

From commit to running cluster

1

Installation

A single stackablectl command bootstraps ArgoCD and the app-of-apps root application.

2

Sync from Git

ArgoCD pulls the manifests and deploys operators and data products automatically.

3

Watch reconciliation

Open the ArgoCD UI to see every component turn healthy and in-sync.

4

Make a change

Edit a manifest in Git; ArgoCD detects the diff and rolls it out.

5

Roll back

Revert the commit and the cluster returns to its previous state.

Ready to run it yourself?

Put your whole data platform under version control and let ArgoCD keep the cluster in sync.