Two interesting blog posts on Azure Functions performance and scale:

🏎️ From 0 to 32000 requests per second in 7 seconds and

🔥 Improving cold start time

img

This is interesting, because the idea of serverless compute is very compelling, until you run into challenges. Obviously there isn’t a machine running your code all-the-time, because that would be costly for either the customer and/or the cloud provider. The challenge then is, how do you make deployments and scaling “feel” like it is?

Thanks for reading! :-)