Rate limiting is crucial for protecting federated GraphQL APIs from abuse and DDoS attacks. However, traditional rate limiting solutions don't work well with federated GraphQL APIs due to the complex nature of GraphQL queries. Implementing rate limiting at the Edge can lead to imprecise results, while implementing it within Microservices can create organizational challenges. Instead, Cosmo Router uses Redis to store rate limit state and implements a simple rate limiting strategy based on the number of Subgraph requests within a given time period. This approach is transparent to clients, fast and efficient, and protects APIs from abuse while allowing legitimate requests to pass.