Looking Back on 2015
In January of 2015, Chris and I sat down and decided it was time to commit to Sentry (no pun intended). We opened our first office here in San Francisco, hired the best people we knew, and set out to take Sentry to an entirely new level. Let’s take a look at what happened in 2015.
We’ve put a lot of focus on our SaaS service the last few years, and 2015 we really saw that paying off. It’s pretty amazing what our small team has accomplished.
- More than 15,000,000,000 — thats 15 billion — errors reported to sentry.io.
- The average event weighed in at 35KB. More than 500 terabytes of error data.
- We surpassed 50,000 user accounts.
- Rounding out the end of the year saw the public launch of Sentry 8, our biggest release to date.
Our infrastructure grew, but not as much as we estimated (which is great!):
- From 33 machines up to 43 in three datacenters.
- While only a 30% increase in machine count, that represented a 50% increase in cost as we scaled up.
- Riak was the largest hardware footprint in early 2015, and today it has been surpassed by Redis, in addition to being nearly leveled out with Postgres.
- 848 unique commits nearly evenly distributed, with Python slightly winning.
- 111 combined contributors, with Python and Ruby being neck-in-neck for most unique.
The biggest change of 2015 was us going heads down on Sentry. We had a pretty big challenge ahead of us when the year started.
- We grew the team from two to eight full-time employees — every single one helping to better your experience.
- Sentry’s office finally moved out of an apartment and into our first real space in San Francisco.
- The team in Austria added a new member, Adrian Ronacher — born July 2015.
This will be a big year for Sentry. We’re anticipating growing the team to 20 by the end of the year. On the product side our initial focus is on first-class support for mobile platforms as well as the Unity game engine. Beyond that, we’re going to tackle the signal-to-noise problem that exists with alerts. It’s a big challenge, but we’re pretty excited to solve it. Finally, expect to see a lot more functionality around general workflows with your team, including VCS support — release logs, git blame, and commit-to-resolve.