What do people use as /health
endpoint for Route53 when you have active-active architecture with microservices (on EKS)? Do you usually stand up one “health” pod per region or you try to aggregate somehow health checks from all running services within the cluster?
We had tried at one point calling /health and having it call it’s dependencies’ /health etc - but became a bit of a nightmare, and if one service was down there was already an alert getting generated… so we ended up with just having a health check hit each service and use knowledge to understand the downstream impacts etc
But when you configure Route53 health check you can only specify one /health
endpoint. So, there is no way to check each service individually
That’s why I was thinking to have all microservices report health to one place and then that places exposes one /health
endpoint
Or just have one simple K8s pod/service with /health
endpoint