We’re building a self service centrally managed Kubernetes platform on AWS that would allow different product teams in our organization to seamlessly run their containerized workloads without needing to worry about service operations(that includes Infra management, Secret rotation, Observability, Incident Management, Capacity Planning etc). My team’s goal is to keep as many operational tasks as possible away from the development teams and make it cloud agnostic.
Currently, this solution is deployed in the single AWS region but couple of our services runs in a different AWS region and they want to onboard to this platform. Hence, we want design the multi-region architecture.
Any guidance on design and operational aspects to consider for managing multi-region SaaS Platform.
02/09/2022, 7:47 PM
There are couple of considerations :
1. Are all sites / region symmetric in terms of size or configuration ?
2. How do you handle contious update across sites ?
3. handling partial failure
4. How do you know weather a site is in good condition ?
My point is that managing operation across multiple sites is the same as managing a single site multiple times.
I can expand in this if you want privately