Oct 29, 2018

Diffy finds potential bugs in your service using running instances of your new code and your old code side by side

What is Diffy?

Diffy finds potential bugs in your service using running instances of your new code and your old code side by side. Diffy behaves as a proxy and multicasts whatever requests it receives to each of the running instances. It then compares the responses, and reports any regressions that may surface from those comparisons. The premise for Diffy is that if two implementations of the service return "similar" responses for a sufficiently large and diverse set of requests, then the two implementations can be treated as equivalent and the newer implementation is regression-free.

How does Diffy work?

Diffy acts as a proxy that accepts requests drawn from any source that you provide and multicasts each of those requests to three different service instances:

A candidate instance running your new code
A primary instance running your last known-good code
A secondary instance running the same known-good code as the primary instance
As Diffy receives a request, it is multicast and sent to your candidate, primary, and secondary instances. When those services send responses back, Diffy compares those responses and looks for two things:

Raw differences observed between the candidate and primary instances.
Non-deterministic noise observed between the primary and secondary instances. Since both of these instances are running known-good code, you should expect responses to be in agreement. If not, your service may have non-deterministic behavior, which is to be expected.
Diffy measures how often primary and secondary disagree with each other vs. how often primary and candidate disagree with each other. If these measurements are roughly the same, then Diffy determines that there is nothing wrong and that the error can be ignored.