Notes on a Wiring Language
LibEMP has a level of complication that I wish to avoid when it comes to operational setup and deployment. The primary pain comes from the configuration, layout, and wiring of Monitors to Buffers and Buffers to Sinks on a per host basis. Part of my idea to alleviate some of this pain comes with the definition of a Wiring language.
Goals of a Wiring Language:
- Each Component can be configured for their startup.
- Each Monitor can choose which Buffer to push to (unless LibEMP forces single semantics Buffer only).
- Each Sink can choose which Buffer it reads from (unless LibEMP forces single semantics Buffer only).
- Each LibEMP Node can define which Monitor/Buffer/Sink starts up.
- A configuration can be loaded dynamically on an already running Node (thus Nodes can start blank).
- The configuration script can be version controlled.
- The configuration script is simple to learn/use and automatically generate.
//TODO: Look into patching into some of the configuration management systems out there. I would rather use one of them rather than roll our own.
