Backend Engineer · 2023
Redis-Backed Rate Limiter
A distributed rate limiter built on Redis to protect core APIs from abuse and traffic spikes without adding meaningful latency.
40%
Latency reduction on protected endpoints
99.99%
Uptime during peak traffic events
10M+
Requests/day handled
The problem
Core public APIs were vulnerable to traffic spikes and abusive clients, causing cascading slowdowns across dependent services.
The approach
Designed a token-bucket rate limiter backed by Redis, using Lua scripts for atomic check-and-decrement operations, deployed as a shared library across services with per-client and per-route configuration.
Full write-up of the architecture, trade-offs considered (Redis vs. in-memory vs. API gateway-level limiting), and the rollout process goes here. This is editable rich text in TinaCMS.