The SRX platform, by default, generates SNMP traps in cases of high routing engine (RE) CPU usage, but no traps are generated for data plane CPU/SPU utilization.

This article provides a method to generate SNMP traps when the data plane CPU utilization on the Service Processing Unit (SPU) is high.

The SRX platform, by default, generates SNMP traps in cases of high Routing Engine (RE) CPU usage, but no traps are generated for data plane CPU/SPU utilization.

The SRX platform generates the following messages when data plane CPU utilization crosses the threshold conditions. The same messages can be used to generate traps by using event-policy .

PERF_MON: RTPERF_CPU_THRESHOLD_EXCEEDED: FPC 6 PIC 1 CPU utilization exceeds threshold, current value=86
PERF_MON: RTPERF_CPU_USAGE_OK: FPC 6 PIC 1 CPU utilization returns to normal, current value=77
PERF_MON: RTPERF_CPU_THRESHOLD_EXCEEDED: FPC 6 PIC 1 CPU utilization exceeds threshold, current value=85
PERF_MON: RTPERF_CPU_USAGE_OK: FPC 6 PIC 1 CPU utilization returns to normal, current value=73

Here is the configuration.

To generate traps when utilization crosses the threshold:

set event-options policy <Policy 1>  events RTPERF_CPU_THRESHOLD_EXCEEDED
set event-options policy <Policy 1> then raise-trap

To generate traps when utilization comes down below the threshold:

set event-options policy <Policy 2> events RTPERF_CPU_USAGE_OK
set event-options policy <Policy 2> then raise-trap

2020-07-08: Article reviewed for accuracy; no changes required; article still valid