Skip to main content

google_compute_autoscaler resource

Syntax

A google_compute_autoscaler is used to test a Google Autoscaler resource

Beta Resource

This resource has beta fields available. To retrieve these fields, include beta: true in the constructor for the resource

Examples

describe google_compute_autoscaler(project: 'chef-gcp-inspec', zone: 'zone', name: 'inspec-gcp-autoscaler') do
  it { should exist }

  its('target') { should match /\/inspec-gcp-igm$/ }
  its('autoscaling_policy.max_num_replicas') { should eq '5' }
  its('autoscaling_policy.min_num_replicas') { should eq '1' }
  its('autoscaling_policy.cool_down_period_sec') { should eq '60' }
  its('autoscaling_policy.cpu_utilization.utilization_target') { should eq '0.5' }
end

describe google_compute_autoscaler(project: 'chef-gcp-inspec', zone: 'zone', name: 'nonexistent') do
  it { should_not exist }
end

Properties

Properties that can be accessed from the google_compute_autoscaler resource:

id
Unique identifier for the resource.
creation_timestamp
Creation timestamp in RFC3339 text format.
name
Name of the resource. The name must be 1-63 characters long and match the regular expression [a-z]([-a-z0-9]*[a-z0-9])? which means the first character must be a lowercase letter, and all following characters must be a dash, lowercase letter, or digit, except the last character, which cannot be a dash.
description
An optional description of this resource.
autoscaling_policy
The configuration parameters for the autoscaling algorithm. You can define one or more of the policies for an autoscaler: cpuUtilization, customMetricUtilizations, and loadBalancingUtilization. If none of these are specified, the default will be to autoscale based on cpuUtilization to 0.6 or 60%.
min_num_replicas
The minimum number of replicas that the autoscaler can scale down to. This cannot be less than 0. If not provided, autoscaler will choose a default value depending on maximum number of instances allowed.
max_num_replicas
The maximum number of instances that the autoscaler can scale up to. This is required when creating or updating an autoscaler. The maximum number of replicas should not be lower than minimal number of replicas.
cool_down_period_sec
The number of seconds that the autoscaler should wait before it starts collecting information from a new instance. This prevents the autoscaler from collecting information when the instance is initializing, during which the collected usage would not be reliable. The default time autoscaler waits is 60 seconds. Virtual machine initialization times might vary because of numerous factors. We recommend that you test how long an instance may take to initialize. To do this, create an instance and time the startup process.
cpu_utilization
Defines the CPU utilization policy that allows the autoscaler to scale based on the average CPU utilization of a managed instance group.
utilization_target
The target CPU utilization that the autoscaler should maintain. Must be a float value in the range (0, 1]. If not specified, the default is 0.6. If the CPU level is below the target utilization, the autoscaler scales down the number of instances until it reaches the minimum number of instances you specified or until the average CPU of your instances reaches the target utilization. If the average CPU is above the target utilization, the autoscaler scales up until it reaches the maximum number of instances you specified or until the average utilization reaches the target utilization.
custom_metric_utilizations
Configuration parameters of autoscaling based on a custom metric.
metric
The identifier (type) of the Stackdriver Monitoring metric. The metric cannot have negative values. The metric must have a value type of INT64 or DOUBLE.
single_instance_assignment
(Beta only) If scaling is based on a per-group metric value that represents the total amount of work to be done or resource usage, set this value to an amount assigned for a single instance of the scaled group. The autoscaler will keep the number of instances proportional to the value of this metric, the metric itself should not change value due to group resizing. For example, a good metric to use with the target is pubsub.googleapis.com/subscription/num_undelivered_messages or a custom metric exporting the total number of requests coming to your instances. A bad example would be a metric exporting an average or median latency, since this value can’t include a chunk assignable to a single instance, it could be better used with utilization_target instead.
utilization_target
The target value of the metric that autoscaler should maintain. This must be a positive value. A utilization metric scales number of virtual machines handling requests to increase or decrease proportionally to the metric. For example, a good metric to use as a utilizationTarget is www.googleapis.com/compute/instance/network/received_bytes_count. The autoscaler will work to keep this value constant for each of the instances.
utilization_target_type
Defines how target utilization value is expressed for a Stackdriver Monitoring metric.

Possible values:

  • GAUGE
  • DELTA_PER_SECOND
  • DELTA_PER_MINUTE
filter
(Beta only) A filter string to be used as the filter string for a Stackdriver Monitoring TimeSeries.list API call. This filter is used to select a specific TimeSeries for the purpose of autoscaling and to determine whether the metric is exporting per-instance or per-group data. You can only use the AND operator for joining selectors. You can only use direct equality comparison operator (=) without any functions for each selector. You can specify the metric in both the filter string and in the metric field. However, if specified in both places, the metric must be identical. The monitored resource type determines what kind of values are expected for the metric. If it is a gce_instance, the autoscaler expects the metric to include a separate TimeSeries for each instance in a group. In such a case, you cannot filter on resource labels. If the resource type is any other value, the autoscaler expects this metric to contain values that apply to the entire autoscaled instance group and resource label filtering can be performed to point autoscaler at the correct TimeSeries to scale upon. This is called a per-group metric for the purpose of autoscaling. If not specified, the type defaults to gce_instance. You should provide a filter that is selective enough to pick just one TimeSeries for the autoscaled group or for each of the instances (if you are using gce_instance resource type). If multiple TimeSeries are returned upon the query execution, the autoscaler will sum their respective values to obtain its scaling value.
load_balancing_utilization
Configuration parameters of autoscaling based on a load balancer.
utilization_target
Fraction of backend capacity utilization (set in HTTP(s) load balancing configuration) that autoscaler should maintain. Must be a positive float value. If not defined, the default is 0.8.
target
URL of the managed instance group that this autoscaler will scale.
zone
URL of the zone where the instance group resides.

GCP Permissions

Ensure the Compute Engine API is enabled for the current project.

Edit this page on GitHub.

Thank you for your feedback!

×









Search Results