Tag
4 posts
In-process limiters count per instance, so ten replicas allow ten times the limit. A shared counter in Redis makes the limit global, at the cost of a network hop and a failure policy.
A limiter is a function of time, so tests should control time. Assert the boundaries: the last allowed request, the first rejected one and the moment capacity returns.
A token bucket allows bursts up to a capacity and a steady refill rate. The whole algorithm is a timestamp and a float, and the async part is waiting for the next token.
Limit requests per client in Axum with a small middleware, then see when to switch to tower_governor or a Redis counter. Includes fixed windows, token buckets and 429 responses.