Upbot

Distributed uptime monitoring. Check your sites from 15+ global regions, get alerted the moment they go down, and watch latency trends in real time.

Upbot pings your endpoints from Cloudflare's edge network across regions like Dublin, Sydney, Frankfurt, São Paulo, Dubai and more, fans the work out through a Redis Streams queue, and streams results back to a live dashboard. It's built as a Turborepo monorepo so the scheduler, queue, edge checkers, API and UI each stay independent and independently deployable.

Why it's built this way

A single server pinging every site on a timer doesn't scale and can't tell you whether your users in another continent can reach you. Upbot separates scheduling, queueing, edge execution, and persistence so each part can scale on its own and a slow region never blocks the others. The full reasoning and tradeoffs live in ARCHITECTURE.md.

Architecture

                  ┌──────────────┐
                  │   web (UI)   │  Next.js 15 + React 19 + Recharts
                  └──────▲───────┘
                         │ REST
                  ┌──────┴───────┐
   schedules      │    server    │  Express API: auth (JWT/OTP/bcrypt),
   checks         │  /api/v1/*   │  websites, alerts, checks, profiles
      │           └──────▲───────┘
┌─────┴──────┐          │ Prisma
│   pusher   │   ┌──────┴───────┐
│ scheduler/ │   │    store     │  Prisma + PostgreSQL (schema + seed)
│ producer   │   └──────▲───────┘
└─────┬──────┘          │
      │ xAdd            │ persist results
      ▼                 │
┌──────────────┐   ┌────┴─────┐      ┌─────────────────────┐
│ Redis Streams│──▶│  worker  │─────▶│  cfworkers (edge)   │
│   (queue)    │   │ consumer │ POST │ Cloudflare Workers   │
└──────────────┘   │  group   │◀─────│ multi-region pings   │
                   └──────────┘ res  └─────────────────────┘

Data flow: pusher reads the websites due for a check, expands each into per-region jobs based on the user's plan (getRegionsForPlan), and xAdds them onto a Redis Stream. worker consumes the stream as a consumer group (so multiple workers share the load and stuck messages are reclaimed via claimPendingMessages), batches up to 50 jobs, and dispatches them to the Cloudflare Worker. cfworkers runs the actual HTTP checks from each requested edge region, returns status + latency, and the worker persists results through Prisma. The server exposes that data over a versioned REST API, and web renders dashboards and charts.

Monorepo layout

Path	What it is
apps/pusher	Scheduler/producer that selects regions per plan and enqueues checks onto Redis Streams
apps/worker	Consumer-group worker that batches jobs, calls the edge workers, and persists results
apps/cfworkers	Cloudflare Workers that perform the multi-region HTTP checks (Vitest tests)
apps/server	Express REST API: auth, profile, website, alert, check, team (v1)
apps/web	Next.js 15 dashboard (React 19, Recharts, Tailwind 4)
apps/client	Next.js 15 front-end app
apps/tests	Cross-cutting integration tests
packages/redis-streams	Redis Streams client: xAddBulk, xReadGroup, xAckBulk, claimPendingMessages
packages/store	Prisma client, schema, and seed data
packages/ui	Shared React component library
packages/eslint-config, packages/typescript-config	Shared tooling configs

Tech stack

Languages/runtime: TypeScript, Bun, Node >=18 Backend: Express, Prisma + PostgreSQL, Redis Streams, Cloudflare Workers Frontend: Next.js 15, React 19, Tailwind CSS 4, Recharts Tooling: Turborepo, Bun workspaces, ESLint, Prettier, Vitest

Getting started

Prerequisites: Bun >= 1.1, PostgreSQL, Redis, and (for edge checks) a Cloudflare account + Wrangler.

# 1. Install
bun install

# 2. Configure env. Copy the example in each app that has one
cp apps/server/.env.example apps/server/.env
cp apps/worker/.env.example apps/worker/.env
cp apps/pusher/.env.example apps/pusher/.env
cp packages/store/.env.example packages/store/.env

# 3. Set up the database
cd packages/store
bunx prisma generate
bunx prisma migrate dev      # or: psql < seed.sql

# 4. Run everything in dev
bun run dev

Key environment variables

App	Variable	Purpose
store	DATABASE_URL	PostgreSQL connection string
worker	WORKER_URL	Cloudflare Worker endpoint that runs the checks
worker	MAX_BATCH_SIZE	Max jobs pulled per stream read (default 50)
worker	WEBSITE_TIMEOUT	Per-check timeout in ms
server	PORT	API port (default 8080)

Useful scripts

bun run dev          # turbo run dev (all apps)
bun run build        # turbo run build
bun run lint         # turbo run lint
bun run check-types  # turbo run check-types
bun run format       # prettier

Roadmap

• Team/collaboration routes (scaffolded, currently disabled)
• Status pages and incident timelines
• Webhook / Slack / email alert channels

License

See repository for license details.