Brume

Brume is a self-hosted PaaS: point it at a Git repository containing a Dockerfile, set your environment variables and domain, and Brume builds the image and deploys your app onto a k3s cluster with automatic routing, TLS, autoscaling and observability.

It runs on a single server or a fleet — the orchestration model is the same from 1 to N nodes.

Pillars

• Résilience — multi-replica by default, health probes, pod disruption budgets, node spreading, zero-downtime rollouts with automatic rollback, replicated storage (Longhorn), backups.
• Auto-scaling — per-app Horizontal Pod Autoscaler (CPU/RAM) with configurable min/max replicas.
• Observabilité & Alerting — Prometheus metrics, Grafana dashboards, Loki logs, and Alertmanager alerts (CrashLoop, OOM, failed deploys, saturation, expiring certificates) via email / webhook / Slack.

Stack

Layer	Choice
Orchestration	k3s
Control-plane	Go (chi) + PostgreSQL
Image build	Kaniko (in-cluster Job) from a Dockerfile + Trivy scan
Registry	private in-cluster registry
Storage	Longhorn (replicated, node-independent volumes)
Routing / TLS	Traefik (bundled with k3s) + cert-manager
Managed addons	Postgres / MySQL / Redis per environment
Observability	Prometheus + Grafana + Loki + Alertmanager
Web UI	Go + templ + HTMX

See docs/ARCHITECTURE.md for the full design and docs/ROADMAP.md for the phased delivery plan.

Development

make db-up     # start a local Postgres
make dev       # run the control-plane against it
make test      # unit tests

The control-plane boots without a database in "smoke mode" (P0); /readyz reflects database availability.

Endpoints (P0)

Path	Purpose
GET /healthz	liveness
GET /readyz	readiness (checks DB)
GET /version	build metadata
GET /metrics	Prometheus metrics
GET /api/v1/ping	API smoke check

Status

Early development. Current phase: P0 — Foundations (see roadmap).