Skip to content

PLATFORM_STATUS.md

Latest commit

 

History

History
133 lines (105 loc) · 5.73 KB

PLATFORM_STATUS.md

File metadata and controls

133 lines (105 loc) · 5.73 KB

Platform Status — willbracken.com

Shared context for Claude and Gemini. Updated as work progresses. Last major update: April 7, 2026.

✅ LIVE AND WORKING

Cloudflare Workers (all deployed, observability on)

Worker URL Purpose
ai-router ai-router.will-i-bracken.workers.dev Tiered AI routing to 5 agents
site-assistant site-assistant.will-i-bracken.workers.dev Public chat + contact form
cf-agent cf-agent.will-i-bracken.workers.dev CF management (Gemini-powered)
cf-monitor cf-monitor.will-i-bracken.workers.dev Daily free health monitoring
analytics-dashboard analytics-dashboard.will-i-bracken.workers.dev D1 + Supabase metrics
r2-explorer r2-explorer.will-i-bracken.workers.dev R2 file browser + graph
agent-code agent-code.will-i-bracken.workers.dev GPT-4o code specialist
agent-research agent-research.will-i-bracken.workers.dev GPT-4o research synthesis
agent-data agent-data.will-i-bracken.workers.dev SQL + data analysis
agent-writing agent-writing.will-i-bracken.workers.dev Claude Sonnet writing
agent-gemini agent-gemini.will-i-bracken.workers.dev Gemini Flash/Pro/Ultra direct
security-txt security-txt-worker.will-i-bracken.workers.dev Security policy

Infrastructure

  • Cloudflare Vectorize: email-search, knowledge-base, platform-docs (768-dim cosine)
  • D1 willbracken-observability: requests, ai_usage, security_events, nodes, edges, hourly_patterns
  • Supabase: ai_usage_log (live data), ai_daily_costs + ai_provider_summary views
  • Analytics Engine: platform_analytics dataset (all workers write to it)
  • MacBook tunnel: ollama-macbook.willbracken.com — 28 models (local + cloud wrappers)
  • n8n: n8n.willbracken.com ✔ BACK ONLINE (Railway US East migration complete)

Code in repos

  • db_adapter.py — Snowflake → Postgres adapter for Streamlit apps
  • Railway Dockerfile + railway.toml for Streamlit deployment
  • Family tasks CF Pages form (BrackenW3/Cloudflare/apps/family-tasks)
  • Landing page force graph (BrackenW3/Cloudflare/apps/willbracken-landing)
  • 6 n8n workflow JSON files (need import to n8n — see below)
  • Standard new-repo GitHub Actions template

⏳ JUST RESOLVED

  • n8n back online — Railway migrated to US East (same region as Supabase)
  • DB fix applied — DB_POSTGRESDB_POOL_SIZE=2 (prevents connection pool exhaustion)
  • GitHub Actions optimized — disabled 3 duplicate AI reviewers, added caching

📋 NEEDS ACTION (quick wins)

Task Est. Time Notes
Import n8n workflows 5 min n8n is UP: Settings → Import from URL (see GEMINI.md for URLs)
Supabase: 2 missing tables 2 min SQL Editor: CREATE workflow_logs + infra_events
Atlassian account 5 min atlassian.com — personal, separate from work
Hyperdrive setup 1 min Need actual DB password (placeholder was given)
CF Pages: connect landing 10 min apps/willbracken-landing in BrackenW3/Cloudflare
CF Pages: connect family-tasks 5 min apps/family-tasks in BrackenW3/Cloudflare
Metabase on Railway 5 min metabase/metabase:latest Docker image
Neo4j Aura free account 5 min neo4j.com/cloud/aura (for email graph)
Self-hosted runner on i9 10 min Set GITHUB_RUNNER_TOKEN, docker compose --profile runner

🗓️ BACKLOG (planned)

Email Pipeline (next focus)

  • Armstrong IMAP ingestion (@zoominternet.net, shared family mailbox)
  • 2x Outlook accounts + personal Outlook (flooded, good test case)
  • Spam filtering + unsubscribe automation
  • Tiered storage: R2 (warm <2yr) → Google Drive (cool 2-5yr) → iDrive e2 (cold 5yr+)
  • Vectorize embedding for hot/warm tier (<2yr)
  • MongoDB for email document storage
  • Neo4j graph (Person→Email→Task→Topic→Thread)
  • Email search interface (CF Worker, natural language)
  • Task detection → Jira

Atlassian Integration (after account setup)

  • Jira project: FAM (Project/Task/Step issue types)
  • Confluence: Family Notes space
  • n8n workflows for task capture + Jira sync
  • Sunday weekly digest email

Security / Identity

  • HashiCorp Vault HCP — migrate KV secrets
  • Vault PKI → automate mTLS cert rotation
  • Azure Entra SSO for Streamlit apps (personal tenant)
  • Desktop tunnel: Windows i9 cloudflared setup

Observability

  • Analytics Engine anomaly detection (n8n workflow)
  • Metabase dashboards: AI costs, worker metrics, Jira stats
  • Cost alerts: if daily spend > threshold → notify

🚫 DEFERRED (conscious decisions)

  • GCP services (not needed at scale)
  • AWS compute (Railway covers it)
  • Codespaces (too expensive)
  • Tableau (Metabase covers use case)
  • go-tools in docker-compose (low priority)

Repositories

Repo Purpose Main branch
BrackenW3/brackenw3.github.io This portfolio site main
BrackenW3/Cloudflare CF Workers, Pages apps, docker-mcp main
BrackenW3/n8n n8n workflows, Railway config main
BrackenW3/VSCode_Folders Cross-platform projects, Streamlit apps main
BrackenW3/WSL_Docker Windows WSL2 workspace main

n8n Workflow Import URLs

Paste these in n8n → Settings → Import → From URL:

https://raw.githubusercontent.com/BrackenW3/n8n/main/workflows/family-task-capture.json
https://raw.githubusercontent.com/BrackenW3/n8n/main/workflows/family-knowledge-archive.json
https://raw.githubusercontent.com/BrackenW3/n8n/main/workflows/family-weekly-digest.json
https://raw.githubusercontent.com/BrackenW3/n8n/main/workflows/email-ingestion-realtime.json
https://raw.githubusercontent.com/BrackenW3/n8n/main/workflows/email-ingestion-batch.json
https://raw.githubusercontent.com/BrackenW3/n8n/main/workflows/ollama-multi-node-router.json