Running PostgreSQL on K8S can be extremely beneficial thanks to advanced resource management, automated deployment and portability. Managing database configurations with ease in accordance with the declarative Infrastructure-as-Code paradigm would make another profit. Better isolation, capabilities ensuring HA and scaling - aren’t those enough to turn anyone into PG+K8S fans?
However, now your DBAs need to master both PostgreSQL and K8S to mitigate the possible risks - and such professionals are rare gems and can be overly expensive. Also, handling persistent storage and data consistency can still be challenging compared to traditional setups. Networking configurations in Kubernetes become intricate, add to this complex backup and recovery strategies and performance overhead, particularly in terms of I/O operations - and you doubt it all very much!
Our PG+K8S debate will give both advocates and sceptics an opportunity to present their viewpoints, and it’s up to you whose opinion to share and support.
Debaters in favor:
Emrah Becer (European Patent Office)
Jonathan Battiato (EDB)
Debaters against:
Matt Cornillon (Google)
Dmitry Fomin (Adyen)
Moderator: Floor Drees
