Personas behavior fp = hash of attribute key set — same fp means identical header surface sent to the API

Behavior fingerprint · credential surface per persona

A behavior fingerprint is a hash of a persona's credential header keys — not values. Two personas with the same key set share the same fingerprint. Adding or removing a header key changes the fingerprint; updating the value alone does not.

fingerprint analysis appears here once personas are created

Device Types each device type sets a request header on every probe request for scenarios using that device dimension

Probe header impact

When a test scenario has a device type assigned, the probe sets User-Agent on every HTTP request it sends for that scenario. This lets Sentinel distinguish device-class anomalies — a mobile UA may receive different API behaviour, response schema, or content than a desktop UA. No device type assigned means the probe sends its default UA.

header impact summary appears here once device types are created

Network Locations each location runs one probe · the probe injects X-Network-Persona into every request

How network identity works

Each network location is served by one probe binary. When the probe executes a test scenario for that location, it injects X-Network-Persona: <value> as an HTTP request header. The target server can use this header to return network-specific responses, which Sentinel then compares against the baseline for that cohort. A scenario with a location assigned is split into a separate baseline per location, so anomalies detected on one network do not suppress alerts on another.