Command Palette

Search for a command to run...

Hectal
Phase 7Intermediate8 of 17 in Redis

Rate Limiting & Distributed Counters

Fixed windows, sliding logs and counters, token and leaky buckets, a production-grade distributed limiter, and counters that survive viral traffic.

Rate limiters and counters are the most common Redis interview designs because they combine everything: the right structure, atomicity with Lua, key design for Cluster, hot keys, and a clear fail-open or fail-closed decision.

You'll implement each algorithm, see exactly where each one lets too many requests through, and then design a multi-tier limiter for 100K requests/sec that keeps working when a Redis node fails.

0/4 · 0%
4 topics ~42 min 11 code blocks & diagrams
Start with the first topic
1
7.1

Fixed Window, Sliding Log, and Sliding Window Counter

A fixed window counts requests per clock minute with INCR + EXPIRE, simple but bursty at window edges. A sliding log stores every request timestamp in a sorted set for exact limits at high memory cost. A sliding window counter blends two fixed windows for a close approximation at fixed-window cost.

10 min 1 diagram 2 code practice

2
7.2

Token Bucket, Leaky Bucket, and GCRA

A token bucket allows bursts up to its capacity and refills at a steady rate; a leaky bucket smooths output to a fixed rate. Both fit in one small Redis hash updated by a Lua script. GCRA achieves the same with a single timestamp.

11 min 1 diagram 2 code practice

3
7.3

Designing a Distributed Rate Limiter

A production limiter combines several scopes (per IP, user, API, tenant), several tiers (free vs premium), atomic Lua decisions, cluster-friendly keys, hot-key handling, and a deliberate choice of what happens when Redis is slow or down.

11 min 1 diagram 2 code practice

4
7.4

Distributed Counters: Single, Sharded, Approximate

A single INCR key is exact and simple until one key gets too hot. Sharded counters spread writes across N keys and sum on read; buffered counters aggregate locally and flush; approximate counters (HyperLogLog, Count-Min) trade exactness for tiny memory. Persist counts that matter to a durable store.

10 min 1 diagram 1 code practice