Our little toad has grown up to be quite the bullfrog. Atticus always hops at the opportunity to catch a new error, but from time to time we’ve experienced floods of incoming errors from a single project, often thousands per minute. We’ve occasionally had to disable such projects, but we’ve never had a policy regarding the number of errors each user could submit. However, we’ve now reached a point where we’ve realized that allowing a single project to use up so many resources is unrealistic given our goal of providing reliable service to all of our customers.
Therefore, we’re rolling out rate limits for all Hoptoad projects. Each Hoptoad plan now has a maximum number of errors allowed per minute for each project. After reaching that threshold, incoming errors for that project will be discarded for the rest of that minute. Errors will be accepted again as soon as that minute is over, and the performance of your application should not be affected. If your account is affected by this change, you’ll know - you’ll be notified by e-mail whenever a project goes over the rate limit for your plan (but only once per day for each project).
This means that we no longer plan on totally disabling projects as we have in the past. It also means that a flood of errors from a single project won’t cause problems for anyone else, and most projects are well below the limit, so few users will experience throttling.
These changes will preserve or improve the level of service for all plans, and put a more effective and fair solution in place for projects with large error traffic spikes.