Skip to main content
Ratelimitly logo
Ratelimitly
Engineering blog

Notes from the edge of scale

Deep-dives on rate limiting, load shedding, performance, and keeping APIs fast under pressure.

Subscribe for new posts
The same 60-request burst enters two limiters: a fresh fixed window admits all 60; a full token bucket spends its 5 tokens, admits 5, and rejects 55. Tokens refill at 1 per second.
Architecture Medium

Why you shouldn’t use Redis as a rate limiter: Part 1 of 2

A tour of the common Redis-based rate limiter implementations — and the correctness and performance traps each one hides.

Read on Medium
Lambda scale-out overloads a fixed-capacity database. Coordinated local shedders return 503 for excess work and target an admitted rate below database capacity so the queue drains. Dashed arrows show shared latency feedback.
Reliability Substack

Auto-Scaling Won’t Save You

The myth of infinite serverless scale — why adding machines doesn’t fix overload, and what to do instead.

Read on Substack

Get new posts in your inbox

Occasional, practical writing on building systems that stay fast under load.

Subscribe