Compare the Paths
Four ways to run Labs64.IO, side by side. If you just want a recommendation: start with the Quickstart, and come back here when you need to decide what comes after it.
At a glance
| Quickstart | One module | Your cluster | Full ecosystem | |
|---|---|---|---|---|
| For | Evaluating | Exploring one service | A team or production | Contributing |
| Runs on | Any Kubernetes cluster | Docker Compose | Your Kubernetes cluster | A local k3d cluster |
| Software comes from | Published images and charts | Built from source | Published images and charts | Built from source |
| Gateway and authentication | Yes, with a demo identity provider | No | Yes, with your identity provider | Yes, with a mock identity provider or Keycloak |
| Data stores | Bundled, in the cluster | Bundled, in Compose | Yours, or bundled | Bundled, in the cluster |
| Services included | AuditFlow, Payment Gateway | AuditFlow | The ones you choose | All of them |
| Time to first request | About 5 minutes | A few minutes | An afternoon for a first environment | One long first build |
| Safe to expose | No | No | Yes, after the go-live checks | No |
What each path needs on your machine
| Tool | Quickstart | One module | Your cluster | Full ecosystem |
|---|---|---|---|---|
kubectl | ✓ | ✓ | ✓ | |
| Helm 4 | ✓ | ✓ | ✓ | |
| Docker | For a local cluster | ✓ | ✓ | |
just | ✓ | ✓ | ||
| JDK 25 and Maven | ✓ | |||
| k3d and Helmfile | ✓ |
The full ecosystem builds images inside a container, so it needs no JDK or Node on your machine. Its dev container brings every tool in the last column.
Which one, then?
flowchart TD
Q{"What do you want to learn?"}
Q -->|"Does this platform do what I need?"| A["Quickstart"]
Q -->|"How does one service behave?"| B["One module locally"]
Q -->|"How do we run it for real?"| C["Your Kubernetes cluster"]
Q -->|"How do I change it?"| D["Full ecosystem from source"]
A common sequence for a team adopting the platform:
- One person runs the Quickstart on a laptop to see whether the platform fits.
- The same person runs one module to learn its configuration in detail, usually AuditFlow and its pipelines.
- The platform team installs it on your cluster, first with bundled data stores, then with managed ones.
- Whoever needs a change the platform does not have yet sets up the full ecosystem and opens a pull request.
Local defaults are not production configuration. The first, second and fourth path use well-known passwords and an identity provider that authenticates nobody. Only the cluster path, with the go-live checks done, is meant to face other people.