8a568932b4
Complete K8s deployment alternative to Docker Compose, targeting
~50 beta users on a Standard_D8s_v5 Azure VM (8 vCPU, 32 GB RAM).
setup-k8s.sh (6 phases):
1. Pre-flight: verify docker phases 1-5 ran, disk/RAM checks
2. Install k3s: Docker runtime, NodePort range 1024-32767
3. Build images: docker compose build + tag as bytelyst/<svc>
4. Config: namespaces, ConfigMap (3 copies), Secrets (JWT + blob keys), Ollama
5. Deploy: infra -> platform -> dashboards -> products (ordered)
6. Health check: 32 endpoints + kubectl pod status
K8s manifests (18 files):
- 4 namespaces (infra, platform, dashboards, products)
- 6 infra (cosmos StatefulSet+PVC, azurite StatefulSet+PVC,
mailpit, loki StatefulSet+PVC, grafana+PVC, ollama external)
- 3 platform (Deployment+Service+NodePort each)
- 2 dashboards (Deployment+Service+NodePort each)
- 10 backends + 9 webs (all with readiness+liveness probes,
resource limits, product-specific NEXT_PUBLIC_* env vars)
Design decisions:
- k3s --docker: reuses existing Docker images, no containerd import
- Same ports as Docker Compose (NodePort with extended range)
- ConfigMap replaces .env.ecosystem, copied to 3 app namespaces
- Blob storage keys injected at deploy time via Secret (not in YAML)
- Cross-namespace DNS: <svc>.<ns>.svc for service discovery
- Ollama as Endpoints+Service pointing to host node IP
- Resource limits: ~19 Gi total, fits in 32 GB with 13 GB headroom
- Teardown: --teardown flag deletes namespaces, keeps k3s
|
||
|---|---|---|
| .. | ||
| docker | ||
| k8s | ||
| README.md | ||
ByteLyst Single-VM Deployment
Deploy the entire ByteLyst ecosystem (30 services, 10 products) on a single Azure VM. Two orchestration approaches — pick one or learn both side by side.
Approaches
docker/ — Docker Compose (Production-ready)
Proven, battle-tested deployment using docker-compose.ecosystem.yml.
Installs everything from scratch on a raw Ubuntu VM in ~20 minutes.
sudo ./docker/setup.sh # Full install
sudo ./docker/setup.sh --resume # Resume after disconnect
/opt/bytelyst/check-health.sh # Verify all 30 services
Use this if: You want reliable deployment now.
k8s/ — Kubernetes via k3s (Learning / Future-ready)
Same 30 services orchestrated by Kubernetes on a single VM using k3s. Builds on the same Docker images — no Dockerfile changes needed.
Use this if: You want to learn K8s with real services, practice kubectl,
and prepare for multi-node scaling later.
Architecture (shared by both approaches)
Raw Ubuntu 24.04 VM (Standard_D8s_v5: 8 vCPU, 32 GB RAM)
├── Ollama (systemd, :11434) ─── local LLM inference
├── Gitea (Docker/:3300) ──────── npm package registry
└── 30 Services
├── Infrastructure (6): cosmos-emulator, azurite, mailpit, loki, grafana, traefik
├── Platform (3): platform-service, extraction-service, mcp-server
├── Dashboards (2): admin-web, tracker-web
├── Backends (10): peakpulse, chronomind, jarvisjr, nomgap, mindlyst,
│ lysnrai, notelett, flowmonk, actiontrail, localmemgpt
└── Web Apps (9): lysnrai-dashboard, chronomind-web, jarvisjr-web, flowmonk-web,
notelett-web, mindlyst-web, nomgap-web, actiontrail-web, localmemgpt-web
Comparison
| Docker Compose | K8s (k3s) | |
|---|---|---|
| Setup time | ~20 min | ~30 min |
| RAM overhead | ~100 MB | ~600 MB |
| Config files | 1 compose + 1 .env | ~30 manifests (or Helm) |
| Scaling | Manual | kubectl scale / HPA |
| Rolling updates | Restart-based | Zero-downtime |
| Resource limits | Basic | Fine-grained per pod |
| Multi-VM ready | Docker Swarm | Native kubectl join |
| Learning value | Low | High (transferable to AKS/EKS/GKE) |