Virtual Service Registry for Serverless Logic Web Tools

Goal: Create an 'in memory/virtual' service registry to allow seamless integration between different projects imported on Serverless Logic Web Tools.

Example use case:

• User imports a sample project A from GitHub that contains a workflow 'Workflow A' he wants to reuse;

• Users start a new project B to create their 'Workflow B' that will consume 'Workflow A' as a fuction. Without this feature, the user would need to register sample project A to a service registry to be able to consume it on project B. 

This step slows down the development cycle because it would require deployment on the service registry. Also, if users need a change on workflow A, it will need a redeploy.

• With a virtual service registry, as soon as the user imports a project inside the web tools, it will automatically register on a 'virtual in memory' service registry and will be automatically available for composition. 

• Also, on each change of 'Workflow A', the Serverless Logic Web Tools will automatically update the registry on the virtual service registry.

As a stretch goal, we will also investigate if it's possible to check workflow dependencies before the deployment. This is useful in cases where in our sample, Workflow B depends on Workflow A, and the user wants to deploy WorkflowB and didn't deploy yet Workflow A (lives only in the virtual service registry).

As soon as the user asks to deploy project B, Web Tools will identify this dependency and ask if the user also wants to deploy Workflow A (the one that B depends on) and provide the proper bindings on Workflow A from the virtual service registry to the deployed one.

Target use cases:

• A serverless workflow that wants to reuse an external sample or pattern as a dependency on authoring, and the user didn't yet publish those samples in their cluster;
• A serverless workflow that wants to consume another project with serverless decisions;
• Allow the user to develop web tools two projects of the serverless workflow at the same time and consume them in authoring mode (without requiring a deployment)