Self-hosted deployment overview
Deploy Scalekit on your own Kubernetes cluster to meet data residency, compliance, and network isolation requirements.
You will run the full Scalekit platform on your own Kubernetes cluster. This helps you keep all authentication data inside your network for data residency, compliance, and isolation from the public internet.
Use self-hosted deployment when your organization requires:
- Data residency: Auth data must remain in a specific region or on-premises location
- Network isolation: The auth service must not be reachable from the public internet
- Compliance: Regulations such as FedRAMP, HIPAA, or internal security policies prohibit use of SaaS auth services
- Air-gapped environments: Your deployment environment has no outbound internet access
How self-hosted deployment works
Section titled “How self-hosted deployment works”Scalekit is deployed on Kubernetes via a Helm chart managed through the Scalekit distribution portal. You configure the deployment with a values.yaml file, apply Kubernetes secrets to your cluster, and connect the cluster to the portal. The portal handles chart delivery, migrations, and upgrades.
Components in a self-hosted deployment
Section titled “Components in a self-hosted deployment”A self-hosted deployment runs as a single Kubernetes Deployment with multiple containers:
| Component | Description |
|---|---|
| Auth service | Core service handling login, token issuance, sessions, SSO, and SCIM |
| Dashboard | Admin web UI for managing your Scalekit instance |
| Flagd | Feature flag sidecar for runtime configuration |
| Webhooks | Webhook delivery service |
| OpenFGA | Fine-grained authorization engine (optional) |
Infrastructure you must provide
Section titled “Infrastructure you must provide”Scalekit does not bundle a database or cache in production. You provision and manage these separately:
| Dependency | Requirement | Notes |
|---|---|---|
| Kubernetes | 1.27 or later | Any managed or self-managed cluster |
| PostgreSQL | 15 or later | Three databases required: scalekit, webhooks, openfga |
| Redis | 6.2 or later | Used for sessions, caching, and job queues |
| SMTP | Any provider | Postmark and SendGrid have first-class support |
Choose your path
Section titled “Choose your path”| Path | When to use |
|---|---|
| Quickstart: Deploy Scalekit on Kubernetes | Get Scalekit running fast using bundled PostgreSQL and Redis. Ideal for evaluation and proof-of-concept. No external databases or kubectl secret commands needed. |
| Full installation | Production deployment with external PostgreSQL and Redis, full secret management, and your choice of ingress controller. |
Next, Quickstart: Deploy Scalekit on Kubernetes will get you a working evaluation instance so you can test before full deployment.