BRANCH YOUR PRODUCTION DATABASE LIKE YOU BRANCH YOUR CODE

Summary from the TLDR Data newsletter

What if Postgres compute was stateless? Imagine a branch that is a metadata operation that runs in seconds, not hours. What if you could also run transactions and analytics on the same copy of data in real time without delays or CDC? One architecture, two problems solved.