GIANT ROBOTS SMASHING INTO OTHER GIANT ROBOTS

Written by thoughtbot

Hoptoad hosting and plan changes

We recently decided to move to a different hosting environment, and we're pleased to announce that Hoptoad has been running very smoothly. The new environment supports the very high write rates that are inherent to an application like Hoptoad, and also affords us higher configurability and flexibility. To put some numbers on it, a couple of months ago we were handling on average 10 thousand requests per minute. That number right now is over 100 thousand. Similarly, our average response time two months ago was around 110ms, whereas now it is averaging at 35 ms. Big wins all around.

We pulled the trigger about a week ago, and we're still tweaking that environment to get the best out of it. We're also doing a lot of performance optimization work on the app itself, but both infrastructure and app performance strategies are topics for another post.

Service changes

Hoptoad

Today we are announcing a couple of changes to the Hoptoad plans.

Plan limits

Based on the usage patterns in Hoptoad, people infrequently look at errors after even a few days, let alone a month. Hoptoad is valuable for daily operations, and developers strive to fix errors in apps as soon as they can. Like with network monitoring software, the usefulness of granular data decreases as time goes by.

The following is somewhat complicated, but we are still going to explain it here so that it is clear. You probably won't even notice that data has been removed.

There are many levels of granularity in the exception data that you see in Hoptoad. At the core there is an error. The error has a bunch of details, like a backtrace, a hash containing the rack environment, parameters, and others. In Hoptoad, many errors get binned together into a group based on their similarity.

Currently, we store all groups and their errors forever, and we delete error details after a month. Thirty days from now, on May 15, we will start deleting errors as well, however you will always know that errors occurred because the group will still be available. In fact, you will still be able to see the error message and class, the file and line number, controller and action, rails environment, and when it occurred last, even after the more granular data is deleted. We will keep error data depending on your plan level:

Egg (Free plan) 5 days
Tadpole 14 days
Heroku Basic 14 days
Toad 21 days
Heroku Plus 21 days
Bullfrog 30 days
Bossfrog 60 days

SSL for all paid plans

We will also start allowing SSL to all paid plans. We believe that SSL is no longer a luxury, so if you are on a paid plan you can enjoy encrypted content, and therefore pqIow DPdFsZiAH Jfz$Wz Bf0jsf08 apijRALPHqaROCKSfw# qWEfqw4#%gG $^b!4t.

We will roll out the data time limits a month from now, but SSL is available to paid plans as of right now.

I'd like to take this opportunity to thank you for your patience while we migrated our environment. We look forward to continue improving the service.