Rate limiter

A simple, distributed rate limiter service built with Go, Redis, Terraform and GCP.

Usage

The service exposes a REST api described in the Openapi document. It exposes the following endpoints:

GET / and GET /healthz perform a health check and returns 200 if the service can connect to Redis.
GET /token/{token_id} will increment the request count for the given token, and will return a response that details the number of tokens remaining, and the number of seconds until reset.

Features

Infrastructure described with terraform, including a bootstrap script, for deploying to GCP.
CI/CD config in cloudbuild.yaml.

Developers

make build builds the binary.
make test runs unit tests. Note that this requires an instance of Redis 4 to be running locally.
make up starts a full local environment with docker-compose.
make deploy runs the cloud init and deployment scripts, powered by Terraform.

Todo

There are a few TODO's in the code base for small code improvements. (grep -r TODO .)
Functions are deployed to public endpoints, without IAM authentication.
Cloud functions have a warm up time that can sometimes be a few seconds. This could be unsuitable in production if consistent latency is important.