
LEGAU
The five-layer industrialised foundation that ensures you never start from scratch.
LEGAU is the shared backbone of every BETC Fullsix digital project: a five-layer architecture (L0 to L4) that pools infrastructure, shared services and reusable business code across all engagements. Each new project inherits a battle-tested foundation and can go live in weeks, not months.
Never from scratch
Layers L0 to L2 are shared across every project. Kubernetes infrastructure, Redis/PostgreSQL/S3/MinIO services, Keycloak SSO - all already running. Client-specific code (L3/L4) is the only genuinely new work per engagement.
Full reversibility
Open code, industry standards (Helm, ArgoCD, Kubernetes) and zero proprietary lock-in. Hand the project over to another agency or take it in-house tomorrow - everything is portable and documented.
Industrialised delivery
CI/CD with semantic-release, automatic versioning and reproducible deployments. One identical pipeline covers the full lifecycle from feature branch to BSO production.
Five-layer architecture
LEGAU organises every project into five independent abstraction layers:
- L0 - Shared infrastructure: Kubernetes on BSO, ArgoCD orchestration, Helm packaging, secrets management via Bitwarden (BWS). Shared across all projects, maintained once.
- L1 - Shared services: PostgreSQL, Redis, S3/MinIO object storage, Keycloak SSO. Every project consumes these services through standardised configurations - no per-client setup.
- L2 - Core package: a reusable library per product line (MUH for headless, VOKAR for conversational, etc.). Semantically versioned, tested and published to the private GitLab registry.
- L3 - Client bundle: client-specific configuration - visual theme, domains, environment variables, ACLs.
- L4 - Project code: the business logic unique to each engagement. The only layer that is genuinely written fresh.
What this means for your CIO
- Declarative GitOps: all infrastructure configuration is version-controlled in Git. ArgoCD reads the manifests - no manual cluster intervention in production.
- Clear branch workflow:
main(stable production, semantic tagsX.Y.Z) /dev(staging, pre-releasesX.Y.Z-pre.N) / feature branches. Identical across all projects, unambiguous. - Automated tests on every merge: code quality, unit tests, e2e, OWASP security, performance - no per-project reconfiguration needed.
- Staging before production: every pre-release tag triggers a deployment to
*.betcfullsix.dev; production is only pushed after explicit human sign-off. - Built-in observability: Loki (logs), Prometheus (metrics), Grafana (dashboards and alerts) are deployed on the cluster and available from the very first project.
- Multi-tenant cost efficiency: layers L0 and L1 are shared across all projects. Each new client only adds the L3/L4 delta, not a full infrastructure stack.