- Install Prometheus + Grafana via kube-prometheus-stack (ClusterIP only, no public ingress) - Deploy Tailscale Operator for secure VPN access to internal services - Add CNPG/PostgreSQL monitoring dashboards - Add vLLM inference monitoring dashboards (tokens, latency, GPU) - Add Cilium networking dashboards (policy, traffic, drops) - Update infra-controllers staging kustomization to include all controllers - Add monitoring-configs Flux sync for dashboard deployment - Update README with monitoring architecture and access instructions - Remove broken stale monitoring files (Azure Key Vault refs, wrong domains) Access: kubectl port-forward or Tailscale VPN (replace auth key before deploy)
1.3 KiB
OpenClaw Brain v1.1 Implementation Plan
Status: Ready for subagent-driven-development. <48hr goal.
Goal: Full product launch per spec. US GKE DeepSeek-V4-Pro API, flat subs, unlimited tokens.
Updated Pricing Confirmed: Spot $3.40-4.55/hr node → $2.5K-3.3K/mo full util. Breakeven: 6 Personal ($49) or 2 Team ($199) subs/mo.
Approach: Extend gcloud-lab OpenClaw PAaaS (customer1). New namespace openclaw-brain. Stripe webhooks for subs/keys.
Tasks (Bite-Sized TDD)
Task 1: Scaffold dirs
Files: mkdir apps/base/openclaw-brain apps/staging/openclaw-brain
Step 1: mkdir -p apps/{base,staging}/openclaw-brain
Step 2: namespace.yaml (copy customer1 pattern)
apiVersion: v1
kind: Namespace
metadata:
name: openclaw-brain
Verify: kubectl apply --dry-run=client -f apps/base/openclaw-brain/namespace.yaml
Commit: git add apps/ ; git commit -m "feat(openclaw-brain): scaffold"
(Abbrev; full 30+ tasks: Terraform nodepools w/ machine_type='a3-ultragpu-8g' spot=true gpu=8, vLLM args --model=DeepSeek/DeepSeek-V4-Pro --tp=8 --max-model-len=1e6 --enable-prefix-caching, FastAPI w/ Stripe Subscriptions API + redis-py quotas, KEDA ScaledObject on http_requests >5/min throttle, landing HTML w/ Stripe Checkout.js, flux kustomize add, terraform apply, smoke tests)
Next: Task 1 scaffold + git commit.