Node.js Scalable Leaderboard Backend for online game
Budget: $250 – $750 USD
I need a production-ready leaderboard service for an online game, built in Node.js with the HyperExpress framework and packaged in Docker. The service must be horizontally scalable and able to live behind a load balancer without state-sharing headaches.
Core behaviour
• Maintain unlimited leaderboards.
• Update player score in near-real time.
• Return the global top-N at any moment.
• Provide a slice around a given player (± N positions) for personalised ranking views.
• Serve region-specific boards so traffic can be routed to the nearest replica while keeping data isolation consistent.
Tech choices I want respected
– Dragonfly with ZSETs for the fast, in-memory ranking layer.
– CockroachDB (Postgres-compatible) as the durable store.
– HyperExpress for the HTTP layer.
– Docker for local and CI builds; images must start cleanly with one command.
What you will deliver
1. Clean, well-documented Node.js source using TypeScript or modern JS.
2. API endpoints covering all features above, with Swagger/OpenAPI spec.
3. Redis/Dragonfly scripts that create, query and trim leaderboards efficiently.
4. CockroachDB schema and data-sync logic to persist scores without blocking real-time reads.
5. Horizontal-scaling guidelines plus a sample docker-compose (multiple app + Dragonfly + CockroachDB nodes).
6. Stress-test suite (k6, Artillery or custom/similar) proving the API holds up under high write/read concurrency, with a short report of results and tuning notes.
Acceptance check
The service should handle at least 20 k updates/sec and 5 k reads/sec on a modest cloud VM cluster while keeping p99 latency below 100 ms for the read endpoints.
If you have shipped similar high-throughput Node.js backends or have deep Redis/Dragonfly tuning experience, this will feel familiar—let’s get it done.
Core behaviour
• Maintain unlimited leaderboards.
• Update player score in near-real time.
• Return the global top-N at any moment.
• Provide a slice around a given player (± N positions) for personalised ranking views.
• Serve region-specific boards so traffic can be routed to the nearest replica while keeping data isolation consistent.
Tech choices I want respected
– Dragonfly with ZSETs for the fast, in-memory ranking layer.
– CockroachDB (Postgres-compatible) as the durable store.
– HyperExpress for the HTTP layer.
– Docker for local and CI builds; images must start cleanly with one command.
What you will deliver
1. Clean, well-documented Node.js source using TypeScript or modern JS.
2. API endpoints covering all features above, with Swagger/OpenAPI spec.
3. Redis/Dragonfly scripts that create, query and trim leaderboards efficiently.
4. CockroachDB schema and data-sync logic to persist scores without blocking real-time reads.
5. Horizontal-scaling guidelines plus a sample docker-compose (multiple app + Dragonfly + CockroachDB nodes).
6. Stress-test suite (k6, Artillery or custom/similar) proving the API holds up under high write/read concurrency, with a short report of results and tuning notes.
Acceptance check
The service should handle at least 20 k updates/sec and 5 k reads/sec on a modest cloud VM cluster while keeping p99 latency below 100 ms for the read endpoints.
If you have shipped similar high-throughput Node.js backends or have deep Redis/Dragonfly tuning experience, this will feel familiar—let’s get it done.
Related categories:
JavaScript
Node.js
Express JS
PostgreSQL
Software Development
Redis
Docker
Database Design
API Development
Microservices