In Prisma SD-WAN, Performance Policies are the primary mechanism used to define the expected quality of experience for specific applications. Unlike traditional monitoring that relies solely on "up/down" interface states, Prisma SD-WAN focuses on the actual health of the application path. An incident is triggered when the system detects a violation of defined service-level agreement (SLA) thresholds, such as excessive latency, jitter, or packet loss, even if the physical link remains active.
When an administrator configures a performance policy, they set specific bounds for these metrics. For example, a VoIP application might have an SLA requiring latency below 150ms and packet loss below 1%. If the ION device detects that the current path (e.g., a broadband circuit) exceeds these limits, it generates a performance incident. This incident serves two purposes: first, it alerts the administrator to the degradation; second, it triggers the Path Selection engine to proactively steer the application traffic to a more suitable "Backup" or "Available" path that currently meets the SLA requirements.
Options B, C, and D represent system-level or network-level events that generate different types of alerts or incidents (System or Network incidents), but they are not the triggers defined within a Performance Policy. Performance policies are specifically concerned with the application's perceived performance across the fabric. By focusing on SLA violations rather than just physical link status, Prisma SD-WAN ensures that business-critical applications remain functional even during "brownout" conditions where a circuit is technically "up" but performing poorly.
