3 ms·
Don't use this unless you know exactly what you are doing. This is not a way to get better performance or scalability in general.
by ltbarcly3 1y ago
Don't use this unless you know exactly what you are doing.
This is not a way to get better performance or scalability in general.
- nijave 1y agoSeems rather niche. You can already improve async write availability with CQRS and a (durable) queue. Systems like Kafka implement this out of the box Seems sort of like a CQRS implementation on top of PG (you're using PG replication as the change queue to loosely separate writes/reads, losing transaction guarantees in the process)