Piglog

Ultra-fast, ultra-clean log aggregation. Grafana capabilities with Linear aesthetics.

Stack

• Frontend: React Router v7, Tailwind CSS, TanStack Query, Recharts
• Backend: Fastify 5, Drizzle ORM, Better Auth, BullMQ
• Database: TimescaleDB (PostgreSQL time-series extension)
• Cache/Queue: Redis
• Deploy: Docker Compose on Coolify VPS

Project Structure

piglog/
├── apps/
│   ├── api/           # Fastify backend (port 3001)
│   └── web/           # React Router frontend (port 3000/5173)
├── packages/
│   ├── contracts/     # Shared Zod schemas + types
│   └── db/            # Drizzle schema, migrations, client
├── ops/docker/        # Dev infrastructure compose
├── compose.prod.yml   # Production stack

Getting Started

1. Environment

cp .env.dev.example .env.dev

.env.dev is gitignored and contains local-only credentials. It's already pre-filled with sensible defaults for localhost development.

2. Infrastructure

npm run dev:infra

This starts TimescaleDB (pg18-oss) and Redis in Docker with persistent volumes.

3. Database Setup

npm install
npm run db:migrate

This runs all .sql migrations in order and tracks them in a _migrations table.

4. Run Dev Servers

You'll need 3 terminal sessions:

# Terminal 1 - API (port 3001)
npm run dev:api

# Terminal 2 - Background workers (alerts, webhooks)
npm run dev:worker

# Terminal 3 - Web dev server (port 5173)
npm run dev:web

The API and workers automatically load .env.dev via dotenv. The web app proxies /api requests to localhost:3001 and falls back to the same URL for API calls.

Architecture

Ingestion Flow

Client → POST /logs/ingest (API Key)
  → Validate source
  → Batch insert into TimescaleDB hypertable
  → Queue alert evaluation (BullMQ)
  → HTTP 202 Accepted

Query Flow

Browser → GET /logs/query?workspaceId=...&service=...
  → Auth session check
  → Drizzle query with time-based chunk exclusion
  → Return 500 rows max (virtualized on frontend)

Alerting Flow

BullMQ Worker (alert:evaluate)
  → Every minute or per-batch trigger
  → Count logs in time window
  → Compare against alert_rule threshold
  → If triggered: queue webhook notification

TimescaleDB Features Used

• Hypertables: log_entry partitioned by time
• Compression: Enabled after 7 days, segmented by workspace/service/level
• Retention: Automatic cleanup after 90 days
• Chunk Exclusion: Fast time-range queries

Production Deploy

cp .env.example .env.prod
# Fill in production values

docker compose -f compose.prod.yml --env-file .env.prod up -d

Services:

• timescaledb - Database
• redis - Cache & BullMQ broker
• migrate - One-off database migrations
• api - Fastify HTTP API
• worker - BullMQ background workers
• web - React Router SSR/static served

Coolify Notes

For Dockerfile-based deploys on Coolify, run database migrations separately from the API container.

• The API image now starts the HTTP server only.
• Run migrations with the same image using: node packages/db/dist/migrate.js
• In Coolify, use a separate job/service for migrations before deploying the API app.
• Point the API app to port 3001 unless you explicitly set PORT to something else.

Salvaged from Flowpigdev

This project reuses the following patterns from the original repo:

• Container structure (web + api + db + redis)
• Better Auth integration with workspace multi-tenancy
• Billing/Stripe scaffolding
• Notification engine
• Webhook delivery system with retries
• Docker production setup
• React Router framework mode configuration
• Tailwind + dark theme styling

License

Private