Snapshotting Data using PubSub and Datastore for Efficient API Serving

I just couldn't wait until next week to talk about a really cool topic: using PubSub and Datastore to store data snapshots for faster API service. This topic is fun to explore because it brings up some tough problems that often come up in distributed systems, like figuring out the order of messages and dealing with race conditions. In this post, I'll talk about a common solution many people might try first - using the PubSub subscription message ordering feature - and why it doesn't actually solve our problem. I'll also share what I think is a better and faster way to solve this problem. So, let's dive in and learn more about PubSub and Datastore on Google Cloud.

00:30 - Table of Contents
01:37 - Consider this design
05:24 - Scenarios where order of events matter
07:28 - The ordered subscription for PubSub
10:19 - The challenge in analytics operations
12:28 - Solution by combining logic & transaction support
16:15 - Code walk through
21:29 - Summary