🌲Timber.io — Smart Log Analysis for Developers
Last week Timber.io came out of stealth mode, and, as an engineer, I thought I’d write a few words about why I’m so excited about the product (I’m not alone btw). Timber is a smart cloud-based logging platform, launching with support for Ruby, Elixir, and Node apps (with more to come!), which extracts structured data from application logs with the aim of making log analysis, inspection, and debugging as easy and quick as possible.
We met the founders Ben and Zach very early on — they had just quit their jobs at Seatgeek, and Timber was still a few mockups and a dream. We co-led a small pre-seed round alongside our friends at Next View Ventures, as well as Wonder Ventures and Ludlow Ventures. Since then, the team’s obsessive focus on product and user experience has resulted in a beautiful, polished product that is not only a pleasure to use, but also translates into actual time savings for engineering teams at scale.
Working with logs is frustrating, and application logs are probably the most frequently used data source developers rely on for debugging — both during development and in production. Log structure in practice is usually an unfortunate mix of half-written for machines to read and half for humans, and far from optimized for either — so there’s some structure, but relying on it can result in fragile code. Developers spend too much precious time searching through logs and trying to interpret them, time that could be spent building.
The state of log aggregation hasn’t changed much since the founding of Splunk, the legacy incumbent in the space, around 15 years ago. Newer services have launched since then, many with great teams and products, but the prevailing metaphor continues to be search, where the user is trying to find a specific string somewhere in the logs, implying that you already know what you’re looking for.
Timber makes searching fast and easy (table stakes), but it also allows developers to ask much more complex questions because Timber augments the raw log data with user and application context by adding request-time information such as user info and ENV data. Timber is not simply a tool for searching raw logs, but helps correlate that data with a more complete snapshot of what was happening at that moment.
This is a pain point I’ve experienced first-hand in every infrastructure I’ve run and there’s never been an easy solution to reach for in the standard developer toolbox. Certainly some great open-source log management components have emerged, from which you can build an ELK stack, for example, and host your own log management infrastructure. This affords probably the most granular control, but there’s nothing free about hosting and managing it yourself and many organizations choose to spend their engineering resources on their core product and business.
I’ve been watching this space for years and Timber’s logging product is the first I’ve seen that actually understands how developers work and solve problems in practice. With the insight that Timber provides into user and application context, I believe Timber will save developers and organizations significant engineering time that was previously wasted digging through tons of raw log data.
Integrating Timber is intended to be as low friction as possible. Today that gets you hosted logging, robust search, exploration of event data through a beautiful and thoughtful product, and basic threshold-based alerting. If you’re a developer sick of wasting time searching through log data, you should try it, it’s the best thing out there today, and it will only get better.