Valkey

My guest today is Daniel Moran.

He is a maintainer of Fake Redis, a pure Python implementation of the Redis API that began as a way to write tests without standing up a real server, and now works as a drop-in replacement for both Redis Py and Valkey Py, pulling over 5 million downloads a month.

Beyond Fake Redis, he maintains Django task scheduler, built a GitHub actions plugin for JetBrains IDEs, and helps run Django Commons, a community organization for keeping Django packages healthy.

Welcome, Daniel.

So I'm wondering as a downstream maintainer, how closely you were watching that and how did it change, how did it impact Fake Redis basically?

It really, I think it kind of changed the main use case from if before it was for testing purposes, just being replacement to run within the Python process.

I started getting issues of it, like it doesn't implement certain Valkey commands and/or it's like the Valkey responses are different or Dragonfly have like a command which is like not in Redis, so it's like can it be implemented? If before it was in the realm of like, uh, one, two million downloads a month, these days it's already have like more than 40 million downloads a month, so it's quite significant.

And, uh, these days it actually has multiple flavors of running.