AverageServiceLoadScalingTrigger
Represents a scaling policy related to an average load of a metric/resource of a service.
Properties
Name | Type | Required |
---|---|---|
MetricName |
string | Yes |
LowerLoadThreshold |
string (double) | Yes |
UpperLoadThreshold |
string (double) | Yes |
ScaleIntervalInSeconds |
integer (int64) | Yes |
UseOnlyPrimaryLoad |
boolean | Yes |
MetricName
Type: string
Required: Yes
The name of the metric for which usage should be tracked.
LowerLoadThreshold
Type: string (double)
Required: Yes
The lower limit of the load below which a scale in operation should be performed.
UpperLoadThreshold
Type: string (double)
Required: Yes
The upper limit of the load beyond which a scale out operation should be performed.
ScaleIntervalInSeconds
Type: integer (int64)
Required: Yes
InclusiveMaximum: 4294967295
InclusiveMinimum: 0
The period in seconds on which a decision is made whether to scale or not.
UseOnlyPrimaryLoad
Type: boolean
Required: Yes
Flag determines whether only the load of primary replica should be considered for scaling.
If set to true, then trigger will only consider the load of primary replicas of stateful service.
If set to false, trigger will consider load of all replicas.
This parameter cannot be set to true for stateless service.