Scaling Properties

Scaling Configuration Properties describes the properties for configuring SBA Gateway scaling.

Table 1. Scaling Configuration Properties
Property Description
scaling.method The method for scaling the number of instances per verticle.
  • auto — Scale based on available CPUs and scaling factor. The number of instances is ignored.
  • manual — Number of instances is specified. The scaling factor is ignored.
  • singleton — Override all instances to 1. Both instances and the scaling factor are ignored.
The default value is auto.
scaling.components.http.instances The cardinal number of HTTP server instances to be deployed. The default value is 4.
scaling.components.http.factor The scaling factor, calculated as the number of available cores divided by the factor value. For example, a value of 1 means use all available CPUs, and 2 means use half of the available CPUs. The default value is 1.
scaling.components.notify.instances The cardinal number of notify client instances to be deployed. The default value is 1.
scaling.components.notify.factor The scaling factor, calculated as the number of available cores divided by the factor value. For example, a value of 1 means use all available CPUs, and 2 means use half of the available CPUs. The default value is 8.
scaling.components.kafka.instances The cardinal number of Kafka event client instances to be deployed. The default value is 2.The default value is 2.
scaling.components.kafka.factor The scaling factor, calculated as the number of available cores divided by the factor value. For example, a value of 1 means use all available CPUs, and 2 means use half of the available CPUs. The default value is 2.
scaling.components.events.instances The cardinal number of event stream instances to be deployed. The default value is 2.
scaling.components.events.factor The scaling factor, calculated as the number of available cores divided by the factor value. For example, a value of 1 means use all available CPUs, and 2 means use half of the available CPUs. The default value is 2.
scaling.components.amqp.instances The cardinal number of CHF AMQP consumer instances to be deployed. The default value is 1.
scaling.components.amqp.factor The scaling factor, calculated as the number of available cores divided by the factor value. For example, a value of 1 means use all available CPUs, and 2 means use half of the available CPUs. The default value is 8.

The following gateway.yaml excerpt shows the common configuration default values for these properties:

scaling:
  # Settings related to scaling the number of instances per verticle
  # method:
  #   auto -> scale automatically based on available CPUs and scaling factor, 
  #           instances will be ignored
  #   manual -> specify number of instances manually, factor will be ignored
  #   singleton -> override all instances to 1, both instances and factor will be ignored
  # Instances:
  #   Cardinal number of instances to be deployed.
  # Factor:
  #   A scaling factor with the following formula:
  #      Available core / factor
  #   Example - 1 means use all available CPUs, 2 means use half of available CPUs etc.
  method: auto
  components:
    # HTTP server
    http:
      instances: 4
      factor: 1
    # Notify client
    notify:
      instances: 1
      factor: 8
    # Event Stream client
    events:
      instances: 2
      factor: 2
    # Kafka event client
    kafka:
      instances: 2
      factor: 2
    # CHF AMQP consumer
    amqp:
      instances: 1
      factor: 8