What value does Cucumber WiFi bring to it’s customers?
Full stack Wi-Fi platform for small and medium enterprises. Transforming the way people use Wi-Fi.
What do you like about Airbrake?
The support team are always there, always helpful and always answer the questions quickly. That’s a big winner for us. We obviously use Airbrake for the default exception handling that comes built it. That’s great for finding the big bugs in our software. In fact, it’s been a better tool for discovering dead servers than the server monitoring tools we use.
More recently, we’ve been using it to log non-critical errors that wouldn’t throw an exception. Previously, we were just logging these to syslog but that was too much to manage. Now we all get a notification when an event happens to a customer, for example. Our customers love the realtime, proactive support we can offer as a result of using Airbrake.
How do you normally deal with Airbrake exception reports?
Panic panic panic. Breath. Panic. Breath. Repeat.
Setup. Environment, Language, Tools, Etc.
The main api is written in Ruby (Rails). All our data crunching is done with an army of Go workers. We’ve got other applications written in pure c for speed and there’s a diminishing list of NodeJS applications out there.
The front-end is all AngularJS. For datastorage, we use Google’s BigQuery right now – we’re currently processing about a billion entries per month. The cost to run our own infrastructure was too high.
How long have you been using Airbrake?
3 years and counting!
Tell us how your company started?
Cucumber Tony was born out of frustration. We were sick of paying high prices for mediocre, complex networking equipment that needed a degree to operate. We’d been slogging away for 8 years before writing our own platform. Turns out keeping thousand and little WiFi boxes connected around the world is pretty darn hard. Why’s it called Cucumber Tony? No-one really knows but it brings some personality to a rather boring field.
What could Airbrake do better?
I’d like to see a way to mute errors. For example we get a load of non-critical errors from Elasticsearch, usually because we’re indexing or moving the hosts around. Or sometimes one of the Engineers will test something that causes a gazzilion errors in a minute. We don’t really need to see these – after report one, it would be cool to mute further exceptions for N minutes.
Better handling of common errors – it’s hard to navigate through them to find which params they had in common. Right now, it’s down to opening 20 tabs.
Error groups – if we could manually group errors together, that would be cool.
Would you like to add anything else?
Favorite Music to Hack to.
My soundcloud stream, that’s all.
Thank you Cucumber Tony for being an awesome member of Airbrake Bug Tracker family for more than 3 years!