Skip to content

Starform — Architecture & Design Hub

Starform is a developer PaaS — connect a Git repo, push, and it runs on managed Kubernetes with databases, object storage, env vars, observability, and usage-based billing attached. Runs on DigitalOcean — platform on DOKS, customer workloads on self-managed Talos + gVisor — kept cloud-portable by a ports/adapters design. The one principle: "Starbase decides, Shuttle applies" — everything flows through desired state.

Starbase decides · Shuttle applies — everything flows through desired state Source & edge Developer · Git repogit push CloudflareCDN · DNS · WAF · TLS edge DO Load Balancerper region Control-plane VPC · central · its own VPC (peered to each regional VPC) Starbase APIdecides what runs · query broker Starbase Workerbuilds · billing cron Stardeckdashboard DO PostgresStarbase DB DO Registryapp images private VPC peering · telemetry read (1/region) Regional VPC /16 · one per region — customer cluster + regional droplets Customer cluster · self-managed Talos + gVisor · per region Shuttleapplies desired state Customer workloadsgVisor-sandboxed pods Envoy GatewayHTTPRoute · L7 metrics OTel Collectortelemetry agent · DaemonSet Grafana Alloyplatform agent metrics-server+ kube-state-metrics Regional droplets · DO VMs — telemetry + ops/jump OTel gatewaybatches · sole writer ClickHouselogs + metrics · plan-based TTL Shipyardjump droplet · post-MVP External SaaS Depotbuild · remote BuildKit Tigrisobject storage · post-MVP Stripepayments Grafana Cloudplatform self-monitoring push pull desired state · 30s traffic logs + metrics platform

The system at a glance. Two VPCs — a Control-plane VPC (Starbase, Stardeck, Postgres, Registry) that decides what runs, and a Regional VPC /16 per region holding the customer cluster and its off-cluster ClickHouse telemetry droplet — joined by private VPC peering (one per region; separate VPCs even at single-region MVP). Plus external SaaS. The spine: "Starbase decides, Shuttle applies" — Shuttle reconciles every 30s.

Blue = desired-state / telemetry · gray = traffic · dashed = external.

Full topology

  • Overview

    What Starform is, the anatomy of a deploy, system topology, the tech stack, and scaling.

  • Starbase

    The control plane — decides what runs: ports, RBAC, Starforge, billing, desired state, Var Groups.

  • Shuttle

    The per-cluster agent — makes each cluster match desired state, reconciling every 30s.

  • Shipyard

    The per-region machine-lifecycle agent — provisions, upgrades, and repairs the Talos fleet (post-MVP).

  • Kubernetes Runtime

    Self-hosted Kubernetes — customer workloads on Talos + gVisor; the split, isolation, and networking.

  • Stardeck

    The Next.js customer dashboard and Mission Control — the bird's-eye observability view.

  • Networking

    The per-region VPC, traffic path (Cloudflare → LB → Envoy → pod), Gateway API, and LB lifecycle.

  • Observability

    The self-contained build-and-operate guide for the telemetry pipeline — logs, metrics, runbooks.

  • Reference

    The canonical-sources index, glossary, open work, and the FR / SC catalog.

New here? Start with What Starform Is · Anatomy of a Deploy · Canonical Sources.