Get Started

Run one module

AuditFlow is the smallest complete slice of the ecosystem: an API that accepts events, a transformer, and a sink, connected by RabbitMQ. Running it alone is the quickest way to see the publish → route → land pattern that every module follows.

Prerequisites

Docker, Java 25, Maven, and just — all four are required for this tier specifically, because AuditFlow's own justfile runs Maven directly on your host. The canonical, versioned install commands live in one place; this page doesn't repeat them:

Java 25 specifically is enforced by a Maven build plugin — an older JDK on your PATH will fail the build before it gets anywhere near AuditFlow's code.

Commands

git clone https://github.com/Labs64/labs64.io-auditflow.git
cd labs64.io-auditflow
just up

How long it takes

First run is dominated by Maven dependency resolution and container image builds; later runs are substantially faster.

You'll know it worked when…

Publish an event and watch it route:

curl -sS -i -X POST http://localhost:8080/audit/publish \
  -H 'Content-Type: application/json' \
  -d '{"eventType":"demo.event","sourceSystem":"get-started-demo","extra":{"hello":"world"}}'

• The curl command above prints HTTP/1.1 200 and an X-Audit-Event-Id response header — not a 400/403 error body. A response of any kind is not enough; it needs to be that specific status and header.
• Swagger UI at http://localhost:8080/swagger-ui.html loads and lists the AuditFlow API.
• The transformer's docs at http://localhost:8081/docs and the sink's docs at http://localhost:8082/docs both load.
• The RabbitMQ management UI at http://localhost:15673 (guest / guest) shows the event moving through its queues.

Run the whole ecosystem

The workspace repo is the entry point that orchestrates all 12 independent repositories with one justfile: it clones every module, builds every image, and deploys the lot to a local k3d cluster behind the auth gateway.

Prerequisites

Docker and just from Tier 1, plus k3d, Helm, the helm-diff plugin, Helmfile, and kubectl — or skip installing any of it individually and open the repo in the bundled DevContainer instead. Unlike Tier 1, you do not need Java or Maven on your host for this tier: the workspace's just build runs Maven inside a builder container.

Commands

git clone https://github.com/Labs64/labs64.io-workspace.git labs64.io
cd labs64.io
just doctor      # verify tooling before spending time
just clone       # fetch all 12 repositories
just up          # build images, create the k3d cluster, deploy

just doctor is worth running first — it checks every tool above is installed and prints its version, before anything else spends time building.

How long it takes

First run is dominated by Maven dependency resolution and container image builds — now across every cloned module, plus the initial k3d cluster bring-up; later runs are substantially faster.

You'll know it worked when…

• http://gateway.localhost responds, instead of connection-refused — the edge is routing.
• The aggregated API docs served by the api-docs chart list endpoints from more than one module, not just one.
• kubectl get pods against the local cluster shows a running pod for each of the five deployable modules — auditflow, auth-gateway, checkout, payment-gateway, customer-portal — not one for every cloned repository (most of the 12 don't ship a workload at all).

Deploy to your own cluster

Every module ships as a published Helm chart. If you already run a Kubernetes cluster — locally, on AWS, or on your own infrastructure — you don't need the workspace repo at all; add the chart repository and install what you need.

Prerequisites

Helm 3.x and kubectl configured against the cluster you intend to deploy to (see the same Prerequisites table for exact versions).

Commands

helm repo add labs64io https://labs64.github.io/labs64.io-helm-charts
helm repo update
helm search repo labs64io

For AWS QA/Staging/Prod sizing, or bringing your own infrastructure, don't improvise values from scratch — the deployment modes are already documented:

How long it takes

Adding and searching the chart repository only fetches a small index file over the network. Installing a chart against your own cluster is bounded by that cluster's own image pulls and startup, not by anything on this page.

You'll know it worked when…

• helm search repo labs64io lists chart names and versions, instead of returning an empty result.
• Those are the same chart versions you can pass to helm install / helm diff upgrade against your cluster.