Alarms provide information about a system's operational condition, which an operator may need to act upon.
MPE or MRA devices generate Policy Server alarms based on the evaluation of component states and external factors. The servers communicate with each other in a cluster. Each server has a database with merge capabilities to replicate the alarm states to the CMP database. This information is shown on the KPI dashboard or in detailed CMP reports.
As alarms and events are raised on an application or the platform, the SNMP subsystem issues a corresponding trap.
Alarms and Events have the following differences:
- Alarms:
- Are issued when a Fault is detected
- Are latched until the Fault is removed (that is, they are explicitly set and cleared)
- Have a Severity: Critical, Major, Minor
- Will cause a trap
- Events:
- Are issued when a Condition is detected (not a Fault)
- Are not latched (that is they are not explicitly set or cleared)
- Do not have a Severity (the Severity is actually INFO)
- Might cause a trap
Separate traps are sent upon raising an alarm and upon clearing an alarm.
Application traps contain the following variable bindings in addition to the
sysOpTime and
trapID fields:
- comcolAlarmSrcNode - The node that originated the alarm
- comcolAlarmNumber - The OID of the alarm and trap
- comcolAlarmInstance - An instance is used when the trap is for a physical device such as disk1, or connection diameterPeer 10.15.22.232:33119
- comcolAlarmSeverity - Severity of the alarm: Critical (1), Major (2), Minor (3), Info (4), Clear (5)
- comcolAlarmText - A text object that defines the trap
- comcolAlarmInfo - An extended text field that adds information to the trap text
- comcolAlarmGroup - The group from which the trap originated (such as PCRF or QP)
Refer to the Policy Management Troubleshooting Reference for more information about Policy Server alarms and traps.