Need to somehow rate limit Open AI calls

Someone could in theory submit a very high number of PRs and make you use up your tokens with OpenAI - need something you can set up to only allow N number of PRs per time period as a circuit breaker (note: this is why for, right now, PRs against this repo itself are not using the webhook after I had it live for testing... once rate limiting is in, will bring it back, sickos)