Honeycomb Beelines Verify Fast Time to Value at hCaptcha
5 minute read

About
hCaptcha is a drop-in replacement for reCAPTCHA that earns website owners money and helps companies get their data labeled.
Environment
- Python, JavaScript
- Kubernetes
- microservices (~20)
- Redis
- PostgreSQL
Goals
An essential part of the hCaptcha service is a JavaScript library that the customer includes within the code of their website. Because the code runs in the customer’s environment, it was critical that the hCaptcha Customer Success team be alerted to problems with captcha sessions immediately and be able to ship fallback captchas, then quickly turn to examining trace data and further investigate the source of the issue—potentially within the customer’s inline code or the hCaptcha service—to understand how to proceed with a resolution.
What They Needed
- Fast, automatic instrumentation of their codebase with events, timings, and quality traces that provide immediate value
- An observability service that allowed them to respond to issues and drill down fast to get the data needed to correctly diagnose and resolve problems
Honeycomb @ hCaptcha
The team at hCaptcha got started with Honeycomb Tracing early on, but recently began getting even more value after installing the Honeycomb Beeline for Python.
The Beelines include insightful and valuable traces by default, they’re built by people who know what is useful. The magic that is there makes sense.
Recently, monitoring alerted the hCaptcha team that the service overall was becoming slow, and that some queries were timing out. By following requests across the system using traces provided by the Honeycomb Beeline, they were able to determine that database connections were not being automatically closed as expected.
The visibility we get within our Kubernetes
Operational responsiveness is just part of the benefit the team is already getting from using Honeycomb. When hCaptcha’s engineering team delivers a new version, they use Honeycomb to watch user experience—if there’s a problem, they can immediately ship a fallback captcha, then fix the problem and/or contact the customer to get it fixed.
Every time a user is shown a captcha, the hCaptcha service sends events and traces to Honeycomb about the user’s experience with that captcha—whether they complete it or not, retry, how long it takes at various points—so the product team at hCaptcha can model where the friction is in the service and what it looks like. This allows them to plan for and make product decisions, as well as quickly decide what type of response is needed to address an issue.
Every day Honeycomb helps us find misconfigurations so we can reach out to our customers and improve their experience.
And the folks at hCaptcha are optimistic about getting still more value out of Honeycomb:
The best part about Honeycomb is how easy it is to make it useful