I am asking this because you may be having different repos for different smaller services which are again deployed on a separate server. What's the easiest way to keep track of all such repos/servers and deploy them?
You would need a service discovery (service) running somewhere, which should be highly available. Build it on zookeeper/etcd etc, and maintain states and other info.
Marcus Pohorely
CTO/CEO @Pagelanes.com
I can tell you how we do it, since we are currently setting up a completely new stack for our app on AWS. The Services are all different Docker Containers in different Repositories on Github.
We build the Container on Git Push with Wercker and then upload them from Wercker to ECR/Dockerhub.
Then we update ECS with the new service. ECS takes care of the Rest.
In the cluster we have Consul running together with registrator. Consul takes care of knowing wich service-container has been launched on which instance and what`s their IP and port is. When you call a service you need to get this info from consul and then directly call it.
Done Deal.
You can exchange ECS with every other Scheduler, Wercker with Jenkins or anything else and AWS with any other provider as long as everything is connected within a VPC and Consul gets the services information.
Greets