Alarm

The alarm parameters are used in service scaling and are not included in the service chaining environment files. These parameters define the thresholds used by Contrail to determine when scaling should take place. This set of parameters is only used the service scaling heat template.
The default time configured under the cooldown parameters is intended to allow the firewall enough time to boot up. If you change the cooldown values, leave sufficient time for each new firewall instance to boot up.
Alarm
meter_nameThe metric monitored by Ceilometer and used by contrail to determine when an additional VM-Series firewall should be deployed or brought down. The heat template uses CPU utilization or bytes per second as metrics for service scaling.
cooldown_initialThe amount time Contrail waits before launching a additional service instance after the initial service instance is launched. The default is 1200 seconds.
cooldown_scaleupThe amount of time Contrail waits between launching additional service instance after the first scale up service instance launch. The default is 1200 seconds.
cooldown_scaledownThe amount of time Contrail waits between shutting down additional service instances after the first scale up service instance shut down. The default is 1200 seconds.
period_highThe interval during which the average CPU load is calculated as high before triggering an alarm. The default value is 300 seconds.
period_lowThe interval during which the average CPU load is calculated as low before triggering an alarm. The default value is 300 seconds.
threshold_highThe value of CPU utilization in percentage or bytes per second that Contrail references before launching a scale up event. The default is 40% CPU utilization or 2800 bytes per second.
threshold_lowThe value of CPU utilization in percentage or bytes per second that Contrail references before launching a scale down event. The default is 20% CPU utilization or 12000 bytes per second.

Related Documentation