Configurable Alarm Thresholds and Traps

The Oracle® Enterprise Session Border Controller supports user-configurable threshold crossing alarms. These configurations let you identify system conditions of varying severity which create corresponding alarms of varying severity. You configure an alarm threshold type which indicates the resource to monitor. The available types are:

  • cpu — CPU utilization monitored as a percentage of total CPU capacity
  • cpu-sipd — CPU utilization for the sip process, monitored as a percentage of total CPU capacity
  • cpu-atcpd — CPU utilization for the atcpd process, monitored as a percentage of total CPU capacity
  • cpu-mbcd — CPU utilization for the mbcd process, monitored as a percentage of total CPU capacity
  • cpu-datapath — CPU utilization for forwarding cores, monitored as a percentage of total CPU capacity
  • memory — memory utilization monitored as a percentage of total memory available

    Note:

    When you configure either an alarm-threshold for memory with severity set to critical, or set the heap-threshold option on system-config, the Enterprise SBC will stop processing traffic if the lowest configured value of the two is reached, regardless of how low the value is.

    When triggered, this alarm issues a corresponding SNMP trap. The system checks utilization every 15 seconds after triggering the alarm and issues the trap again if the memory utilization is still breaching the threshold. This can generate a significant number of traps sent to SNMP management systems.

  • memory-datapath — memory utilization monitored as a percentage of total datapath memory available (mbuf)
  • sessions — allowed utilization monitored as a percentage of session capacity
  • space — remaining disk space (configured in conjunction with the volume parameter - see the Storage Expansion Module Monitoring section of the Accounting Guide for more information.)
  • deny-allocation — denied entry utilization monitored as a percentage of reserved, denied entries.
  • reserved-nsep-sessions — allowed utilization for NSEP, monitored as a percentage of session capacity

For the alarm type you create, the Enterprise SBC can monitor for 1 through 3 severity levels as minor, major, and critical. Each of the severities is configured with a corresponding value that triggers that severity. For example the configuration for a CPU alarm that is enacted when CPU usage reaches 50%: 

alarm-threshold
        type                           cpu
        severity                       minor
        value                          50

You may create addition CPU alarms for increasing severities. For example: 

alarm-threshold
        type                           cpu
        severity                       critical
        value                          90

The alarm state is enacted when the resource defined with the type parameter exceeds the value parameter. When the resource drops below the value parameter, the alarm is cleared.

In addition to raising alarms, you can configure the Enterprise SBC to reject calls after a CPU or memory threshold is reached. You configure this in system-config and sip-config load limits. See SMP-Aware Task Load Limiting in Maintenance and Troubleshooting Guide for more information.

SNMP Traps

When a configured alarm threshold is reached, the Oracle® Enterprise Session Border Controller sends an apSysMgmtGroupTrap. This trap contains the resource type and value for the alarm configured in the alarm-threshold configuration element. The trap does not contain information associated with configured severity for that value. 

apSysMgmtGroupTrap 		NOTIFICATION-TYPE
	OBJECTS		{ apSysMgmtTrapType, apSysMgmtTrapValue }
	STATUS	   	current
	DESCRIPTION
	      " The trap will generated if value of the monitoring object
	      exceeds a certain threshold. "
	::= { apSystemManagementNotifications 1 }

When the resource usage retreats below a configured threshold, the Oracle® Enterprise Session Border Controller sends an apSysMgmtGroupClearTrap. 

apSysMgmtGroupClearTrap		 NOTIFICATION-TYPE
	OBJECTS		{ apSysMgmtTrapType }
	STATUS	   	current
	DESCRIPTION
	      " The trap will generated if value of the monitoring object
	      returns to within a certain threshold.  This signifies that
	      an alarm caused by that monitoring object has been cleared. "
	::= { apSystemManagementNotifications 2 }

The alarm and corresponding traps available through the User Configurable Alarm Thresholds functionality are summarized in the following table.

Alarm Severity Cause Actions
CPU minor

major

critical

high CPU usage apSysMgmtGroupTrap sent with

apSysCPUUtil

apSysMgmtTrapValue

memory minor

major

critical

high memory usage apSysMgmtGroupTrap sent with

apSysMemoryUtil

apSysMgmtTrapValue

sessions minor

major

critical

high provisioned usage apSysMgmtGroupTrap sent with

apSysLicenseCapacity

apSysMgmtTrapValue

space minor

major

critical

high HDD usage, per volume apSysMgmtStorageSpaceAvailThresholdTrap

sent with:

apSysMgmtSpaceAvailCurrent

apSysMgmtSpaceAvailMinorThreshold

apSysMgmtSpaceAvailMajorThreshold

apSysMgmtSpaceAvailCriticalThreshold

apSysMgmtPartitionPath

deny allocation minor

major

critical

high usage of denied ACL entries apSysMgmtGroupTrap sent with

apSysCurrentEndptsDenied

apSysMgmtTrapValue

Alarm Thresholds Configuration

To configure alarm thresholds:

  1. Access the alarm-threshold configuration element. 
    ORACLE# configure terminal
ORACLE(configure)# system
ORACLE(system)# system-config
ORACLE(system-config)# alarm-threshold
ORACLE(alarm-threshold)#
  2. type — Enter the type of resource which this alarm monitors. Valid values include:
    • cpu
    • cpu-sipd
    • cpu-atcpd
    • cpu-mbcd
    • cpu-datapath
    • memory
    • memory-datapath
    • sessions
    • space
    • deny-allocation
    • reserved-nsep-sessions
  3. volume — Enter the logical disk volume this alarm monitors (used only when type is set to space).
  4. severity — Set the severity of the threshold. Valid values include:
    • minor
    • major
    • critical
  5. value — Enter the value from 1 to 100, indicating the percentage, which generates an alarm when exceeded.

    Note:

    When type is set to cpu-datapath or memory-datapath, you must use a number greater than 30.
  6. Save and activate your configuration.