How Honeycomb rapidly scaled using CircleCI and AWS
Since adopting CircleCI and AWS, Honeycomb has simplified their entire build, test, and deploy process, leading to major improvements in speed, scale, and cost-savings.
Founded in 2016 as an observability tool, Honeycomb.io helps modern engineering, DevOps, and site reliability teams operate more efficiently. Honeycomb enables technical teams to understand what’s happening in production by observing, analyzing, and distilling production data, and using those lessons to improve the development process. Among their customers are leading brands like Slack, Optimizely, and Vanguard.
Cost-effective, efficient infrastructure required
As Honeycomb looked to complete their technical stack, they searched for top-tier infrastructure tools that would enable them to ship code quickly and cost-effectively with minimal internal resources. The Honeycomb team was only interested in solutions that could scale rapidly so they could grow their customer base faster than the costs associated with running their service.
Build, test, and deploy, simplified
Honeycomb decided on a complementary hardware and software solution that combines offerings from AWS and CircleCI. They moved from Intel x86 to AWS’ Graviton2 processors (based on Arm architecture), and to simplify the build, test, deploy process on Graviton2 instances, they selected CircleCI’s Arm compute as their CI/CD service. Since using AWS’ Graviton2 and CircleCI, Honeycomb has seen improvements in speed, scale, and cost-savings.
Before adopting Graviton2, Honeycomb had already reduced its build and test cycle time averages from 20-30 minutes down to 10 minutes by leveraging CircleCI’s support for parallelism. Now, using Graviton2 and CircleCI’s Arm resources together, Honeycomb can build multi-architecture Docker images natively, and ship them to production at the same speed they were previously shipping binaries for single architecture.
Deploying Arm-based applications to Graviton2 enables the Honeycomb team to increase their computational capacity up to 5x within their existing infrastructure. With the speed and efficiency of Graviton2 and CircleCI, Honeycomb’s 30 engineers can scale their software development to compete with companies that have 10x the number of engineers.
AWS is Honeycomb’s second-highest expense item after employees. Adopting Graviton2 and CircleCI enabled the team to cut its AWS EC2 compute costs by 50%. In addition, improving build and test speeds by more than 50%, and using a CI/CD managed service has reduced the cost of Honeycomb’s highest expense items: engineers and management.
“When you have the opportunity to increase performance and reduce costs by 50%, selecting Graviton2 and CircleCI is less of a big bet, and more of a ‘why not?’” said Honeycomb’s Principal Developer Advocate, Liz Fong-Jones, of the company’s decision to use these two solutions. Ben Hartshorne, Manager of Platform Engineering, added, “There is no reason not to [use this migration]… It’s cheaper, it’s faster, and everything else just works.”