AssignmentStrategy
Enums · AiDotNet.Enums
Strategy for assigning requests to model versions during A/B testing.
For Beginners
This determines how traffic is distributed between different model versions:
- Random: Each request is randomly assigned based on traffic split percentages.
Use when you want pure statistical randomness.
- Sticky: Users consistently get the same version (based on user ID hash).
Use when you want each user to have a consistent experience across sessions.
- Gradual: Gradually shifts traffic from old to new version over time.
Use when you want to slowly roll out a new version to minimize risk.
Fields
| Field | Summary |
|---|---|
AssignmentStrategy Gradual |
Gradually shift traffic from old to new version over time. |
AssignmentStrategy Random |
Each request randomly assigned based on traffic split. |
AssignmentStrategy Sticky |
Users consistently get the same version (based on user ID hash). |
int value__ |