I'm looking to implement ECS Anywhere to manage containerized applications across over 200 remote locations, each with on-premises servers. Each location's servers are designed to failover only within that location—not to other sites. Given this setup, I'm wondering what the best clustering approach would be. Should I create a separate cluster for each location, or would it be better to group all the locations under a single cluster?
3 Answers
One K8s cluster could work, but considering your failover requirements, I think staying isolated at each location would be smarter. You’d want to avoid adding layers that cross those isolated sites.”},{
Imagine you have a separate manager for each store. That's what managing 200 clusters is like—very safe but maybe a bit tedious. Alternatively, one big cluster is like having just one manager for all stores; if that manager goes down, it affects all of them. A middle-ground approach would be to cluster them into 'districts' with 10 clusters for 20 sites each. This could strike a balance between manageability and safety.
I'd definitely recommend setting up one cluster per location. ECS Anywhere treats clusters as shared management spaces, so if you put all your locations into a single cluster, it increases the risk of something negatively impacting multiple sites all at once. Isolating them by location helps maintain separate failure domains and keeps things easier to manage during local outages. Sure, more clusters means more to manage, but automating that can really help with operations.
Related Questions
Can't Load PhpMyadmin On After Server Update
Redirect www to non-www in Apache Conf
How To Check If Your SSL Cert Is SHA 1
Windows TrackPad Gestures