Back to projects
Distributed systems
QuotaForge
Multi-tenant rate limiter, correct under concurrent load.
Java 21RedisOTel
No live demo yet

Overview
What it does
QuotaForge is a tenant-aware rate-limiting service designed to keep shared APIs responsive and fair under heavy traffic.
Motivation
Why it exists
A single noisy customer should not be able to consume capacity meant for everyone else. QuotaForge makes limits explicit, shared across instances, and safe under concurrent requests.
Capabilities
What to look for
- Tenant-specific quotas and burst limits
- Atomic Redis-backed counters for consistent decisions
- OpenTelemetry instrumentation for traces and metrics
Architecture
How it works
- 1A request is identified by its tenant and the resource it wants to use.
- 2QuotaForge checks and updates the shared Redis counter atomically, so concurrent service instances agree on the result.
- 3The service returns an allow or reject decision together with enough metadata for a client to retry responsibly.
Getting started
How to start it
- 1Clone the project repository and install the Java 21 toolchain.
- 2Start a local Redis instance for shared quota state.
- 3Set the Redis connection and OpenTelemetry settings for your environment.
- 4Start the service with the repository's Gradle wrapper.
Usage
How to use it
- Define a quota for each tenant or API resource.
- Send incoming requests through the limiter before they reach the protected service.
- Use the returned decision and retry metadata to provide clear feedback to clients.
- Inspect telemetry to tune limits and find traffic spikes.