Walkthrough: Upgrading a GKE cluster with zero downtime using blue-green.

Hi everyone! I'm planning my first production GKE cluster upgrade soon and I'm a bit nervous about downtime. I've heard "blue-green" is the way to go for zero downtime, but I'm not entirely clear on the practical steps.

Could someone walk me through a basic, beginner-friendly sequence? I'm especially fuzzy on how traffic switching between the old (blue) and new (green) clusters works in practice. Do I need a load balancer in front of both? Any gotchas to watch out for? 😅

Also, is this approach significantly more expensive since you're running two clusters temporarily? How do you handle database and stateful services during the cutover? Thanks in advance for any guidance!