In this practical hands-on presentation we will do an overview of the Logical Decoding feature and show you how to build high performance memory based read caches using it .
We will demonstrate how to create a configurable application caching service that invalidates and updates it's entries by consuming a PostgreSQL Logical Replication Stream and demonstrate how to handle several very common use-cases for busy web and mobile application backends that need millisecond response times, without sacrificing consistency and improving availability due to reduced DB load.
We will finish with our real-world scenario where we drastically improved critical aggregate queries from several seconds to several milliseconds and reduced our server load by 40% in a 50M transactions per day environment. All this with less than 1000 lines of code.
The following slides have been made available for this session: