Apache Curator - how(best practices)?

Question

Java

Hello, I'm trying to organize the work of many "micro services" that should be in the cloud. I don’t know if I formulated it correctly, since I’m just starting to delve into these concepts ..
The essence of the question (s). Understand what and how to eat.
From the off site and not numerous examples, I'm trying to put everything together.

1. Where can I see an example of a service that would serve as a central registrar for all "micro services"? Is this the most common maven-project that registers a service when, for example, such a call? "host:8080/centralService"?

2. Do I understand correctly that each "micro service" at startup should simply register with the main service, and all other services contact the main one, and it redirects further.

3. What kind of thread tipps...

Answer 1

1. You need to understand that the communication overhead greatly reduces the overall responsiveness of the project, sometimes it is too critical. First of all, you need to deal with Discovery , and IMHO it's better to get down to gradle.
2. This option works in the case of using the Discovery extension - within the application, there are a bunch of services that are registered and subsequent routing takes place depending on the current load. Within the framework of Curator, the raft consensus task algorithm is partially implemented, but not completely, some things usually have to be added by hand, depending on the requirements of the project.
3. First you should deal with ZooKeeper, and then try to cut Curator.