3 minute read
We are thrilled to announce that Surfline has chosen Split to improve phased rollouts and to help their engineering team scale.
You can read the press release here.
Surfline, the market leader in surf forecasting, reporting, live webcasting, photography, videography, and editorial coverage on surfing, is utilizing Split to improve the speed of ongoing feature rollouts and strengthen experimentation capabilities to enhance overall platform efficiency.
Without Split, rollbacks could take a couple hours to a day. And after that, there might still be a lot of code cleanup. Split helped us increase our engineering velocity and code stability.
Drew Newberry, Senior Software Engineering Manager, Surfline
As Surfline grew to become the world’s largest surf forecasting site, with the largest marine camera network in the world, the product and engineering team building and maintaining the site grew as well. The team at Surfline started using various approaches to handle the increasing scale of the team and the enhancements and new products Surfline was rolling out.
The team had been inching closer and closer to a continuous delivery model. Development was done on a main (i.e., trunk). Roughly twice a week release branches were cut off the main and put into staging for quality assurance (QA). If the release candidate passed all tests, it would then be pushed into production. A standard process that many modern engineering teams follow.
The issue was that these deployments were all or nothing. If a release candidate had some important fixes or new features that the team was under pressure to go live with, however, it also introduced some new bugs, there was no way to choose only the desired code updates to be deployed in production.
The team was familiar with feature flagging and considered building an in-house solution. One option was to build something basic that managed feature state through config values. As they were evaluating Split, they saw that not only could it handle their basic needs to separate feature release from code deployment, Split also brought additional functionality. The Split management console made it accessible for product managers to monitor and change the state of a feature rollout easily.
Split separates feature release from code deployment, enabling phased rollouts for major feature updates. And, Split empowers the analytics team to experiment on feature changes anywhere in the codebase.
As a result of deploying Split, Surfline gained a number of key benefits:
- One half of a full-time software engineer’s time regained from not needing to manage rollbacks and code cleanup.
- Minimal engineering time to configure feature flags.
- Split integration to Segment makes it easy to integrate Split impression data with other platforms such as Amazon RedShift.
Read the case study to learn more about how Surfline is using Split.