FAQ:

Is there a free plan?

Honeybadger is free for non-commercial open-source projects — we're actually the error monitoring system of choice for RubyGems.org and Bundler. Other non-profit organizations get a 20% discount. We don't offer a free plan for commercial apps. Frankly, we can't afford it. We're three guys who are feeding our families with the money our awesome customers give us (we aren't venture-backed), so we offer a great service in exchange for your money. It's worked out well so far. :)

What if I need more resources?

No problem. We have larger plans available, and we can customize our pricing to meet your needs. Just shoot us an email, let us know what you need and we'll make it work.

Do you accept purchase orders or checks?

We accept POs and payment by check for users on the Large Annual plan and above.

Are you compliant with HIPAA, SOC-2, or EU SafeHarbor?

Sorry, but for now we're not offering solutions certified compliant with these regulations.

Which languages are supported?

Error monitoring is supported in Ruby, JS, Node, Go, Elixir, Python, PHP, Java, and more -- see our docs for the full list. Uptime monitoring is supported in all platforms.

What do you mean by soft limit?

Here's the deal. We don't have hard limits like Rollbar, and we don't charge you per error like Sentry... when you're having a bad day with a bunch of errors, the last thing you want to worry about is whether your monitoring service is going to cut you off, so we don't do that. That said, we do have to have some kind of limits in place or we'll go bankrupt. :) So what we do is we suggest that you give us more money as you send us more traffic. If it looks like you'll send us more traffic in a month than is allotted to your plan, we'll email you and ask you to upgrade or turn down the volume.

Do you throttle?

We allow for bursts of error traffic without any throttling, but if that burst continues for an extended period of time, we may throttle your project, depending on the total volume of errors we are getting at the time. In other words, if you are sending us 2-3x the traffic of the next-highest sender for more than 10 minutes, we'll limit you to 15 errors per minute until we detect that your flood has subsided. The API will return a 429 response when this happens.

How does retention work?

Most error data gets stale pretty quickly, so we don't keep everything forever. We keep your data around for six months, then we unceremoniously drop one of the table partitions and recycle the disk space.