Distributed system command and control

Question

I have an application that consists of a set of Java and C++ of processes running on one or more systems and need to ability to monitor and control the components as if they were one complete system.

My requirements:

1. runtime statistics (messages per second etc..)
2. log of events
3. control: "service X set rate to 1000 mps" or "all services stop"
4. ambient configuration so all services can start up and discover the working mode (etc..)

Since data plane messaging is 0MQ based, I am currently doing this with 0MQ based messaging and a message distributor to fan out the control messages.

However, I am interested in how others have tackled the problem.

Thanks.

Answer 1

I'm not sure about controlling them, but SNMP is typically used for monitoring of a distributed system like this.

0MQ is awesome however, so there's no reason why one of its configurations can't be used too.