Intercom Speeds up Their Busiest Endpoint by 50%
5 minute read
Intercom makes modern products for sales, marketing, and support to connect with customers and grow faster.
The engineering teams at Intercom knew they had an opportunity to improve the experience of their users by refactoring some aspects of their service, but they didn’t know what factors affected the performance of their code the most. Where should they start, and what work should they prioritize?
What They Needed
- An observability service that allowed them to investigate slowness and drill down to identify how different types of users experienced their services
- A way to evaluate and communicate real data about their environment to support proposals for improvements
Honeycomb @ Intercom
With Honeycomb, Intercom engineers could query and break down across all the dimensions required to understand how different users and types of usage affected the performance of a given endpoint—the type of data, the amount, the particular customer use case—and were able to both identify the portions of the code needing refactoring as well as document concrete examples of how they’d improve performance.
They were then able to roll out this work side-by-side with their existing code, and use Honeycomb to compare the performance of the new code to the existing implementation until they were confident in their improvements and could release it to their entire userbase.
And the result? Major improvements:
This sped up almost all requests to this endpoint, having a larger impact on our larger customers. The median request time dropped 10-15%, and the P99 by around 50%!
(An overview of the difference in time taken between the two approaches.)
Not only has using Honeycomb enabled Intercom to improve their customer experience across the board, but it has also decreased their resource spend:
This in turn lets us turn down capacity as machines can finish work faster.
Happier customers and lower costs? You can get that with Honeycomb.