Building
Strata
A production-grade distributed social platform featuring realtime notifications, activity feeds, Redis caching, BullMQ workers, Dead Letter Queues, monitoring dashboards, and horizontally scalable Socket.IO infrastructure.
Overview
Strata is a backend-first social platform built to explore distributed systems architecture. The project implements a complete social graph, feed generation, realtime notifications, background job processing, Redis caching, failure recovery mechanisms, presence tracking, and operational monitoring. Rather than focusing on UI, the project emphasizes reliability, scalability, and infrastructure design similar to systems used by Twitter, LinkedIn, and Discord.
What Users Can Do
- Create accounts and authenticate using JWT-based sessions.
- Create posts, like posts, and comment on content.
- Follow and unfollow other users to build a personalized social graph.
- View a paginated feed generated from followed users.
- Receive realtime notifications for follows, likes, and comments.
- See unread notification counts update instantly without page refreshes.
- View user presence and online status in realtime.
Why I built this
- To understand how large-scale social platforms architect feeds, notifications, and realtime systems.
- To gain hands-on experience with Redis caching strategies and cache invalidation.
- To implement BullMQ workers and background job processing patterns.
- To learn reliability engineering concepts such as retries, dead letter queues, replay systems, and recovery metrics.
- To build a portfolio project that demonstrates backend engineering beyond standard CRUD applications.
Tech Stack
Next.js
TypeScript
PostgreSQL
Prisma
Redis
BullMQ
Socket.IO
Turborepo
After launch & Impact
- Implemented an event-driven notification pipeline using BullMQ workers and Redis Pub/Sub.
- Built realtime notification delivery using Socket.IO with Redis-backed room synchronization.
- Designed a Dead Letter Queue (DLQ) system with replay capabilities for failed jobs.
- Implemented notification count caching in Redis, reducing repeated database queries.
- Added operational monitoring including queue metrics, worker metrics, health checks, replay history, and recovery rate tracking.
- Integrated Bull Board to provide production-style queue inspection and debugging.
- Implemented online presence tracking and last-seen functionality using Redis.
Future Plans
- Implement fan-out-on-write feed generation using background workers.
- Introduce feed caching and timeline precomputation.
- Add media uploads and content moderation pipelines.
- Deploy multiple API instances behind a load balancer to demonstrate horizontal scaling.
- Build a complete frontend dashboard for notifications, feeds, and system observability.