Dynamic datastream configuration overview
By default, the Adobe Experience Platform Edge Network sends every event arriving on a datastream to all Experience Cloud services you have enabled. Use Dynamic Datastream Configurations to define rules that control which services receive which events and/or which datasets store those events, without changing any client-side SDK code.
Before Dynamic Datastream Configurations, controlling event routing required either maintaining multiple datastreams or adding override logic in your client-side SDK implementation. Dynamic Datastream Configurations move that routing logic server-side into the datastream itself.
What Dynamic Datastream Configurations can do can-do
The following table summarizes the routing actions available to you.
What Dynamic Datastream Configurations cannot do cannot-do
Dynamic datastream configurations are designed for event-level routing. The following actions are not supported.
Rule evaluation model rule-evaluation
Understanding how the Edge Network evaluates rules helps you design configurations that behave predictably.
- First match wins. The Edge Network evaluates rules in the order you define them. When an event matches a rule, the Edge Network applies that rule’s routing configuration and stops evaluating further rules.
- Default fallback. If no rule matches an event, the event follows the default static datastream configuration: the primary event dataset and all enabled services.
- 25ms evaluation budget. All rules in a datastream must evaluate within 25ms total. If evaluation exceeds this budget, the event falls back to the default datastream configuration. Keep rules simple and focused on reliable fields such as
eventType. - Flat expressions only. The system does not support nested logical expressions (containers within containers). If your logic requires nesting, break it into multiple flat rules instead.
For the full list of supported data types, operators, and guardrails, see Create Dynamic Datastream Configurations.
Event value taxonomy event-taxonomy
Before designing rules, classify all event types your implementation sends into one of three categories. This classification directly determines your dataset strategy and rule design.
decisioning.propositionFetch and personalization.requestClassifying your events before you configure rules is the most important planning step. It determines which datasets you need, which events go to which datasets, and how many rules you need to write.
Mutual exclusivity with datastream overrides overrides
edgeConfigOverrides payload.Dynamic datastream configurations and datastream configuration overrides are mutually exclusive per event. When an event carries a client-side override sent via Web SDK sendEvent or configure, the override takes precedence and the Edge Network skips Dynamic Datastream Configuration rules for that event.
Plan your implementation to use one approach or the other per event type. Do not use both. Where possible, use Dynamic Datastream Configurations over client-side overrides. They provide better visibility, traceability, and control.
Next steps
- Review the prerequisites and planning checklist before configuring your first rules.
- Read Dynamic datastream configuration patterns to choose the right dataset strategy.
- Follow the UI steps to create Dynamic Datastream Configurations.