so, how do you actually handle api rate limiting when you're dealing with a ton of traffic?

hey everyone, im working on scaling up our microservices and have hit a common roadblock, rate limiting. we're using an API gateway, but as traffic grows, im wondering about the best strategies for implementing and managing rate limits effectively across many services. should we handle it primarily at the gateway, or should individual services have their own logic? what are your go-to tools or patterns for this, especially for complex scenarios like tiered limits or distributed limiting?