The graph is the basic building block for customer journey execution within Kitewheel.
Each graph represents a series of logical steps that are executed from the Start Node to an end node. A node is either a piece of business logic or an adaptor - the graph executes and each node updates schema and public variable locations to make decisions and to orchestrate events. The graph proceeds from node to node based on the outcome of the node and the Links available to: Error, Conditional or Goto.
A single graph can be executed many times in parallel by the Kitewheel Engine. The Kitewheel Engine will only deploy or visual test valid graphs that are valid.
Graph Validation
A graph is considered valid if it passes all of the following conditions:
- It must have a Start Node (this is automatic - but it is here for completion)
- At least one Return Node
- Each non-return node must have exactly one Goto link and at most one error link
- Each adaptor is valid - all of the necessary connection details and parameters have been specified within the adaptor
- All nodes that require them have their input and output schema locations mapped
- Any sub-graphs must also be valid and not have any listeners
- There is no ghost node in the main graph - it is okay to have ghost nodes, and other disconnected nodes, which are not linked from the Start Node
- Cannot for a cycle - all graphs are directed acyclic graphs
- Any Split Test Node producer or consumer must have a counterpart consumer or producer in the same project
