Struct RedisRateLimiter 

pub struct RedisRateLimiter { /* private fields */ }

RateLimiter over Redis: each allowed request ZADDs now into the key’s sorted set; a request is allowed while the set holds fewer than limit members within the trailing window.

Differences from the Workers Rate Limiting binding (documented,

not papered over)

	Workers Rate Limiting binding	RedisRateLimiter
Algorithm	fixed-window counter (limit + period per binding, configured in wrangler.toml)	sliding window: a request at the window’s edge does not get a fresh full budget immediately
retry_after	the binding exposes none — the port’s Decision::retry_after is always None on Workers	Some: milliseconds until the oldest counted request exits the window, so callers can send an honest Retry-After
Where the limit lives	the binding declaration (per Worker)	config keys RATE_LIMIT_MAX / RATE_LIMIT_PERIOD_SECS, one budget shared by every module — the same “one venture, one budget per subject” the shared binding gives
Scope	counters are per-Cloudflare-edge	one Redis: exact across every process and replica that shares it, and a single point of failure
Key space	the binding’s own	rl:-prefixed in the database shared with RedisKv

Duration-granularity windows are kept whole: the script works in milliseconds.

Implementations

impl RedisRateLimiter

pub fn new(conn: ConnectionManager, limit: u64, window: Duration) -> Self

pub fn from_env(conn: ConnectionManager, config: &dyn Config) -> Self

RATE_LIMIT_MAX (default 100) and RATE_LIMIT_PERIOD_SECS (default 60) from config.

Trait Implementations

impl RateLimiter for RedisRateLimiter

fn limit<'life0, 'life1, 'async_trait>( &'life0 self, key: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<Decision, RateLimitError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,