
A 429 Without Retry-After Is a Bug: How I Rate Limit Django APIs With a Redis Token Bucket
TL;DR: A rate limiter has two jobs: protect everyone else from one noisy client, and tell that client...
Tag archive

TL;DR: A rate limiter has two jobs: protect everyone else from one noisy client, and tell that client...
On load-test day, the P99 for the order details endpoint jumped from 40ms to 900ms. The database was...
A list key quietly became a giant — not because someone wrote a bad loop, but because each iteration paid a full network round trip. The fix is one line, the diagnosis is the real lesson.
An API receives a PDF, asks a model to extract the fields, and returns the result. On your laptop it...
Langfuse ingestion sizing requires accounting for queue recovery, storage, and operations costs; our worked cost model and synthetic OTLP harness support planning, but our blocked local test provides no benchmark results.
Fifth post in the agent sandbox series. A queue makes waiting explicit, and users go from "no capacity" to "you are seventh in line". Reading the code back, I found what I had actually built was not "counters call the next number" but "a floor manage

Why GitLab's Redis Cluster batch helpers still made every caller opt in to cross-slot commands, how I moved that permission into ClusterUtil.batch, and why the issue stayed open on purpose.
Part 7 of 10: Reliable background jobs in Node.js, from zero. By Sandro Dezerio, senior full-stack...
Why single-threaded execution beats multi-threaded databases in latency, and what happens between a TCP packet and physical RAM.
Redis OOM command not allowed: Set maxmemory-policy If Redis is only a cache, run...
Cache-aside with FastAPI and Redis covering TTL, invalidation on write and how to avoid cache stampede, with docker compose and tests.
Valkey vs Redis, tested on Redis 8.10.3 and Valkey 9.1.3. Same speed on my bench, different licences and features, and a migration path that breaks on streams. Here is what I ran, what failed, and which one to pick.