Hey everyone! I'm gearing up to implement a multi-region high availability setup with GKE for a critical application (Identity Platform) in our stack, but I'm new to this kind of deployment. I heard someone mention liqo.io recently, and I came across Google's Fleet option and their Multi Cluster Load Balancer/Infrequent/SVC features. I'm looking for some knowledge-sharing here. Does anyone have recommendations for best practices or personal experiences with this? I'd really appreciate your insights! Thanks!
3 Answers
I might be a bit biased, but GKE Fleet is basically designed for multi-cluster HA. It has other features too, but that's one of the main uses. Just make sure to consider where your state lives, how to sync it across regions, the consistency model you'll need, and how to handle data recovery in case of disasters or partitions.
Are you going multi-region just for HA, or is latency a concern too? A global load balancer with the right health checks could help for front-end traffic. Just keep in mind that managing database replication between regions can get tricky depending on the tech you choose.
Have you considered using a cloud-managed fault-tolerant identity platform for your setup? They might go down occasionally, but they generally offer better uptime than custom-built solutions.
We're trying to avoid vendor lock-in, so we're leaning towards open-source options. We're still exploring our choices though!