GIANT ROBOTS SMASHING INTO OTHER GIANT ROBOTS

Written by thoughtbot

Hoptoad Migration Problems

Yesterday, we pulled the trigger on our new datacenter for Hoptoad. The good news is that the response times are half to a third as long as they were previously, so on that front it was exactly what we wanted. We’re currently pushing arount 50000 requests per minute according to NewRelic RPM, and we’re responding to them in an average of 33ms.

The bad news, unfortunately, is that the data migration did not go according to plan. Because of a synchronization issue, some people were able to errors from other people’s accounts for about an hour and a half.

Because of the amount of data coupled with the rate at which new data was incoming, we found that the only option we could take was to wipe out the details. This did two things: reset the synchronization and it prevented people from seeing unauthorized data.

The data we removed were the details for errors (that is, the backtrace, params, session, and rack environment; the unix ENV is not sent). The errors themselves are still in Hoptoad, but ones that arrived between approximately 4:30pm and 11:30pm Eastern will not have details along with them. Any new errors you get for those same problems will contain the full data.

We apologize greatly for any inconvenience this has caused, and our offer to refund a month of service still stands. To claim your refund, or to report any issue whatsoever that may have arisen because of this migration, please visit our help site and we’ll help as soon as we can.