Why replicate/reproduce user requests?
Rerun Replicate extends the "rerun" button to instantly replicate any real customer/user request.
There's a couple of major use-cases:
Goal 1: Reproduce user issues instantly.
The main purpose of Rerun Replicate is being able to easily reproduce user issues.
When you get a ticket, you just go to the endpoint, find that particular request, and you've instantly got a local reproduction of the issue.
Successfully reproducing a user's exact state often needs more than just the same exact request:
- Auth
- Safely accessing production data
Goal 2: Run realistic tests (including edge-cases).
You've made a change. Looks good locally. But how will it work in the real world? With real data?
You can use Rerun Replicate to sample real requests, and see how it runs in a variety of states.
Goal 3: Understand your code.
Sometimes you're looking at logs thinking "how did we even get here"?
With Rerun Replicate, you can just replicate a given request locally and step through with breakpoints etc. to understand what's going on.