Hacker News Reader

Zappa author here - thank you for your support!

I love zappa, there's also chalice: https://github.com/awslabs/chalice

Have you used both? How would you compare/contrast them?

Is there a big overhead in terms of latency? If I understand correctly you essentially have to start a server for each request.

~20ms, and there is no parallel burden, so it's usually faster in medium to high traffic scenarios.

Not the case - it runs with all dependencies cached if its called multiple times. Inconsistent calls will incur the delay though.