debezium VS temporal_tables

They manage data in the application layer and your original data stays where it is. This way data consistency is no longer an issue as it was with streaming databases. You can use Change Data Capture (CDC) services like Debezium by directly connecting to your primary database, doing computational work, and saving the result back or sending real-time data to output streams.

Both of these articles mention a key player, Debezium. In fact, Debezium has had a place in the modern infrastructure. Let's use a diagram to understand why.

This might be overkill, but you could use an extra tool like https://debezium.io to capture logs about all creates, updates, and deletes in your table

If you're considering a relational database for an event-driven architecture, check out Debezium. It lets you stream changes to relational databases, and subscribe to change events.

Debezium

https://debezium.io/ comes to mind as an open source product, but there are a gazillion of these tools out there.

Contrary to what it sounds, it’s quite straightforward. The database changes are captured via Debezium and published as events on Apache Kafka. ClickHouse consumes those changes in partial order by Kafka Engine. Real-time and eventually consistent.

There is also the temporal_tables extension.

[0] https://github.com/arkhipov/temporal_tables

- https://github.com/arkhipov/temporal_tables

I haven't used any of these but I work on https://xtdb.com which is also implementing SQL:2011's temporal features :)

It seems like PG has this extension here anyone ever use it?

You could look at that or ask me questions about it (disclaimer, I am the author). Also there is https://github.com/arkhipov/temporal_tables/

All good. Note there is also https://github.com/arkhipov/temporal_tables/ (which is also type 4 as a postgres extension - pretty similar to what ebean orm is doing)

The Temporal Tables Postgres extension works well. https://github.com/arkhipov/temporal_tables

If you're doing this across regions, you really really should reconsider. If you're doing it in the same data center you might be able to get away with it (but then I'm not sure why you're doing it in the first place, if the system fits in one DC then you probably can just scale up). It might be worth considering a sharded & passively combined approach -- i.e. every country has it's own data, and there's some huge public schema which consists of all the data that is drip fed in to materialized views or tables at regular intervals. You could also combine this with temporal_tables to get a very delayed but theoretically time-consistent (well, aside from clock skew across regions of course...) view of your DB to query... Really depends on the use case.

One of postgres's most underrated features. RLS is amazing, can be unseen/basically work silently if your programming language-side tools are good enough, and is documented well (like everything else):

https://www.postgresql.org/docs/current/ddl-rowsecurity.html

But the power of PG is that it doesn't stop there, if you combine this with a plugin like temporal_tables and you can segment by user and time:

https://github.com/arkhipov/temporal_tables

All of this mostly unknown to the thing that's accessing the DB. If that's not enough for you, why not add some auditing with pgaudit:

https://www.pgaudit.org/#section_three

I think it might not actually be hyperbole to say that Postgres is the greatest RDBMS database that has ever existed.