In production environments, users will ultimately need services that are guaranteed to stay online. These services will be prefixed with HA for high availability, but there are some challenges to solve.
We need to accomplish the following programmatically with the kick server.
Setup an ELB associated with the cluster's VPC
Using fleetctl, trigger the launch of N instances of a service specified via a *@.service template (@ signifies template to fleetctl, [name]@[id].service)
Wire these services into the ELB. Is this a master-slave configuration or load-splitting?
Establish a way to detect instance health. Is this done by AWS or does the kick server have to measure it?
In production environments, users will ultimately need services that are guaranteed to stay online. These services will be prefixed with HA for high availability, but there are some challenges to solve.
We need to accomplish the following programmatically with the kick server.