Case · Backend · Rust
rate-gate: a request rate limiter
An HTTP service in Rust: one POST /v1/check call says whether an action may pass. Two algorithms: a token bucket on integer milli-tokens (bursts) and a sliding window over two counters (strict limits). JSON rules chosen by key prefix, Prometheus metrics, a state snapshot, a demo page. State lives in the memory of a single instance: replicas do not share it.
- Rust
- axum
- tokio
- proptest
- musl
- distroless
- Prometheus
- Docker
- GitHub Actions
01 · How it looks
Screenshots
02 · About
What it is and why
An HTTP service in Rust: one POST /v1/check call says whether an action may pass. Two algorithms: a token bucket on integer milli-tokens (bursts) and a sliding window over two counters (strict limits). JSON rules chosen by key prefix, Prometheus metrics, a state snapshot, a demo page. State lives in the memory of a single instance: replicas do not share it.
Limitations
A single instance: state is not shared between replicas.
The sliding window is approximate (a weighted sum of two windows).
The snapshot is not transactional: a crash loses up to 30 s of counters.
The project runs on a small VPS with demo data: no redundancy or backups.
What is inside
- A token bucket without floats: milli-tokens, cost above 1, retry_after_ms and reset_ms computed
- Property tests (proptest): over any period the bucket grants at most capacity + refill * elapsed
- 20 tests; an 8.2 MB image, 3.8 MiB RAM idle and 3.9 MiB after 200 requests
- Container uid 10001, read-only, cap-drop ALL, CSP script-src 'self'
03 · How to run
Run it locally
git clone <repository on request> rate-gate
cd rate-gate
cargo run --release04 · Code
The repository is private
The source is in a private repository. Message me on Telegram and I will open access.
Request the code