Scaling based on schedules

Schedule-based autoscaling lets you improve the availability of your workloads by scheduling capacity ahead of anticipated load. If you run your workload on a managed instance group (MIG), you can schedule a required number of virtual machine (VM) instances for recurring load patterns as well as one-off events. Use scaling schedules if your workload takes a long time to initialize and you want to scale out in advance of anticipated load spikes.

This document describes how you can create, list, edit, disable, re-enable, and delete scaling schedules for an existing MIG. For more information about MIGs and autoscaling, see Creating managed instance groups and Autoscaling groups of instances.

Before you begin

  • Read about autoscaler fundamentals.
  • If you haven't already, then set up authentication. Authentication is the process by which your identity is verified for access to Google Cloud services and APIs. To run code or samples from a local development environment, you can authenticate to Compute Engine by selecting one of the following options:

    Select the tab for how you plan to use the samples on this page:

    Console

    When you use the Google Cloud console to access Google Cloud services and APIs, you don't need to set up authentication.

    gcloud

    1. Install the Google Cloud CLI, then initialize it by running the following command: 

      gcloud init
    2. Set a default region and zone.

    Terraform

    To use the Terraform samples on this page in a local development environment, install and initialize the gcloud CLI, and then set up Application Default Credentials with your user credentials.

    1. Install the Google Cloud CLI.

    2. To initialize the gcloud CLI, run the following command: 

      gcloud init

    3. If you're using a local shell, then create local authentication credentials for your user account: 

      gcloud auth application-default login

      You don't need to do this if you're using Cloud Shell.

    For more information, see Set up authentication for a local development environment.

    REST

    To use the REST API samples on this page in a local development environment, you use the credentials you provide to the gcloud CLI.

      Install the Google Cloud CLI, then initialize it by running the following command: 

      gcloud init

    For more information, see Authenticate for using REST in the Google Cloud authentication documentation.

Limitations

Scaling schedules are restricted by the limitations for all autoscalers as well as the following limitations:

  • You can have up to 128 scaling schedules per MIG. To mitigate this limit, delete scaling schedules that have an OBSOLETE status and that you do not plan to run again.
  • The minimum duration for scaling schedules is 5 minutes.

Creating a scaling schedule

You can create up to 128 scaling schedules per MIG. For more information, see Scaling schedule settings.

The following instructions explain how to create a scaling schedule for a MIG.

Console

  1. In the Google Cloud console, go to the Instance groups page.

    Go to Instance groups

  2. Click the name of a MIG from the list.

  3. Click Edit.

  4. If no autoscaling configuration exists:

    1. Under Autoscaling, click Configure autoscaling.
    2. Under Autoscaling mode, select On: add and remove instances to the group to enable autoscaling. If you want to scale your MIG based only on schedules, delete the default CPU utilization metric after you add the schedules.

  5. For each scaling schedule that you want to add:

    1. Expand the Autoscaling schedules section, click Manage schedules, then click Create schedule.
    2. In the Create scaling schedule pane, enter a Name.
    3. Optional: Enter a Description.
    4. Enter the number of Minimum required instances that this schedule provides when it is active.
    5. Specify the start time and recurrence of your scaling schedule either by using the default interface or, if you want to configure a schedule with a more complex start time and recurrence, by using a cron expression.
      • Default interface
        1. In the Start time field, type or click to select a start time.
        2. In the Recurrence field, select how often the schedule repeats. If you select Every week or Every month, use the additional drop-down menu to select which days of the week or month the schedule starts.
      • Cron expression
        1. To enable, click the Use CRON expression toggle.
        2. Enter a CRON expression.

    6. In the Time zone field, select a time zone.

    7. In the Duration and Unit of time fields, enter a duration and select a corresponding unit of time.

    8. Click Save. The Scaling schedules pane opens.

    9. Optional: You can create another scaling schedule by clicking Create schedule.

  6. When you are finished creating schedules, click Done.

  7. To close the Instance groups page, click Save.

gcloud

To add a scaling schedule to a MIG that does not have an existing autoscaling policy, use the gcloud compute instance-groups managed set-autoscaling command.

gcloud compute instance-groups managed set-autoscaling MIG_NAME \
    [--min-num-replicas=MIN_NUM_REPLICAS] \
    --max-num-replicas=MAX_NUM_REPLICAS \
    --set-schedule=SCHEDULE_NAME \
    --schedule-cron="CRON_EXPRESSION" \
    --schedule-duration-sec=DURATION \
    [--schedule-time-zone="TIME_ZONE"] \
    --schedule-min-required-replicas=MIN_REQ_REPLICAS \
    [--schedule-description="DESCRIPTION"] \
    [--zone=ZONE | --region=REGION]

To add a scaling schedule to a MIG with an existing autoscaling policy, use the gcloud compute instance-groups managed update-autoscaling command.

gcloud compute instance-groups managed update-autoscaling MIG_NAME \
    [--min-num-replicas=MIN_NUM_REPLICAS] \
    [--max-num-replicas=MAX_NUM_REPLICAS] \
    --set-schedule=SCHEDULE_NAME \
    --schedule-cron="CRON_EXPRESSION" \
    --schedule-duration-sec=DURATION \
    [--schedule-time-zone="TIME_ZONE"] \
    --schedule-min-required-replicas=MIN_REQ_REPLICAS \
    [--schedule-description="DESCRIPTION"] \
    [--zone=ZONE | --region=REGION]

Replace the following:

  • MIG_NAME: the name of a MIG with an existing autoscaling policy.
  • MIN_NUM_REPLICAS: Optional: the minimum number of instances required for this MIG. Google recommends setting this value to the minimum number of instances that you need when no scaling schedule is active. When you set this value to 0 and configure autoscaling with schedules or Cloud Monitoring metrics with single time series per group, your MIG can scale in to 0 VMs. Scaling in to 0 VMs is not possible when the policy has other signals, which require data from individual VM—for example, CPU utilization.
  • MAX_NUM_REPLICAS: the maximum number of instances that can ever be provided for this MIG. Optional for a MIG with an existing autoscaling policy.
  • SCHEDULE_NAME: the name of the new scaling schedule.
  • CRON_EXPRESSION: the start time and recurrence of this schedule formatted as a cron expression.
  • DURATION: the duration, in seconds, that this schedule is active.
  • TIME_ZONE: Optional: the IANA time zone for the schedule's start time—for example, Europe/Paris. The default value is UTC.
  • MIN_REQ_REPLICAS: the number of minimum required instances that this schedule provides when it is active.
  • DESCRIPTION: Optional: a description of the new scaling schedule.
  • ZONE or REGION: Optional: the zone or region where your MIG is located.

You can create schedules that run once or that periodically repeat.

Repeating schedule

This example command creates a schedule that repeats. This schedule specifies that you want your MIG to contain at least 10 VM instances on every Monday, Tuesday, Wednesday, Thursday, and Friday from 8:30 AM to 5 PM Coordinated Universal Time (UTC). This command also uses the --min-num-replicas and --max-num-replicas flags to update the minimum number of instances to 0 and the maximum number of instances to 30 for all the autoscaling signals of the selected MIG. This command assumes you have an existing MIG named example-mig with an existing autoscaling policy.

gcloud compute instance-groups managed update-autoscaling example-mig \
    --min-num-replicas=0 \
    --max-num-replicas=30 \
    --set-schedule=workday-capacity \
    --schedule-cron="30 8 * * Mon-Fri" \
    --schedule-duration-sec=30600 \
    --schedule-min-required-replicas=10 \
    --schedule-description="Have at least 10 VMs every Monday through Friday from 8:30 AM to 5 PM UTC"

Onetime schedule

This example command creates a schedule that runs once. This schedule specifies that you want your MIG to contain at least 30 VM instances all day on January 30, 2030 for the America/New_York time zone. This command assumes you have an existing MIG named example-mig that is located in us-east1-b and has an existing autoscaling policy.

gcloud compute instance-groups managed update-autoscaling example-mig \
    --set-schedule=example-onetime-schedule \
    --schedule-cron="0 0 30 1 * 2030" \
    --schedule-duration-sec=86400 \
    --schedule-time-zone="America/New_York" \
    --schedule-min-required-replicas=30 \
    --schedule-description="Schedule a minimum of 30 VMs all day for January 30, 2030" \
    --zone=us-east1-b

Terraform

To add a scaling schedule to a MIG, use the google_compute_autoscaler resource.

resource "google_compute_autoscaler" "default" {
  provider = google-beta
  name     = "my-autoscaler"
  zone     = "us-central1-f"
  target   = google_compute_instance_group_manager.default.id

  autoscaling_policy {
    max_replicas    = 5
    min_replicas    = 1
    cooldown_period = 60

    scaling_schedules {
      name                  = "every-weekday-morning"
      description           = "Increase to 2 every weekday at 7AM for 12 hours."
      min_required_replicas = 2
      schedule              = "0 7 * * MON-FRI"
      time_zone             = "America/New_York"
      duration_sec          = 43200
    }
  }
}

To learn how to apply or remove a Terraform configuration, see Basic Terraform commands.

REST

To create an autoscaler with scaling schedules for a MIG, use the autoscalers.insert method for a zonal MIG or the regionAutoscalers.insert method for a regional MIG.

Make the following call to create a schedule for a zonal MIG that does not have an existing autoscaling policy:

POST https://compute.googleapis.com/compute/v1/projects/PROJECT/zones/ZONE/autoscalers/

{
  "name": "AUTOSCALER_NAME",
  "target": "https://www.googleapis.com/compute/v1/projects/PROJECT/zones/ZONE/instanceGroupManagers/MIG_NAME",
  "autoscalingPolicy": {
    "minNumReplicas": MIN_NUM_REPLICAS,
    "maxNumReplicas": MAX_NUM_REPLICAS,
    "scalingSchedules": {
      "SCHEDULE_NAME": {
        "minRequiredReplicas": MIN_REQ_REPLICAS,
        "schedule": "CRON_EXPRESSION",
        "timeZone": "TIME_ZONE",
        "durationSec": DURATION,
        "description": "DESCRIPTION"
      },
      ...
    }
  }
}

To create schedules for a MIG that has an existing autoscaling policy, use the autoscalers.patch method for a zonal MIG or the regionAutoscalers.patch method for a regional MIG.

For example, make the following call to create a schedule for a zonal MIG that has an existing autoscaling policy:

PATCH https://compute.googleapis.com/compute/v1/projects/PROJECT/zones/ZONE/autoscalers?autoscaler=AUTOSCALER_NAME

{
  "autoscalingPolicy": {
    "minNumReplicas": MIN_NUM_REPLICAS,
    "maxNumReplicas": MAX_NUM_REPLICAS,
    "scalingSchedules": {
      "SCHEDULE_NAME": {
        "minRequiredReplicas": MIN_REQ_REPLICAS,
        "schedule": "CRON_EXPRESSION",
        "timeZone": "TIME_ZONE",
        "durationSec": DURATION,
        "description": "DESCRIPTION"
      },
      ...
    }
  }
}

Replace the following:

  • PROJECT: your project id.
  • ZONE: the zone where your MIG is located.
  • AUTOSCALER_NAME: the name of an existing autoscaler. This is often the same as the name of the MIG.
  • MIN_NUM_REPLICAS: Optional: the minimum number of instances required for this MIG. Google recommends setting this value to the minimum number of instances that you need when no scaling schedule is active. When you set this value to 0 and configure autoscaling with schedules or Cloud Monitoring metrics with single time series per group, your MIG can scale in to 0 VMs. Scaling in to 0 VMs is not possible when the policy has other signals, which require data from individual VM—for example, CPU utilization.
  • MAX_NUM_REPLICAS: the maximum number of instances that can ever be provided for this MIG. Optional for a MIG with an existing autoscaling policy.
  • SCHEDULE_NAME: the name of the new scaling schedule.
  • MIN_REQ_REPLICAS: the minimum required instances that this schedule provides when it is active.
  • CRON_EXPRESSION: the start time and recurrence of this schedule formatted as a cron expression.
  • TIME_ZONE: Optional: the IANA time zone for the schedule's start time—for example, Europe/Paris. The default value is UTC.
  • DURATION: the duration, in seconds, that this schedule is active.
  • DESCRIPTION: Optional: a description of the new scaling schedule.

You can create schedules that run once or that periodically repeat.

Repeating schedule

This example API call creates a schedule that repeats. This schedule specifies that you want a zonal MIG to contain at least 10 VM instances on every Monday, Tuesday, Wednesday, Thursday, and Friday from 8:30 AM to 5 PM Coordinated Universal Time (UTC). This API call also uses the minNumReplicas and maxNumReplicas fields to update the minimum number of instances to 0 and the maximum number of instances to 30 for all the autoscaling signals of the selected MIG. This API call assumes that you have an existing project named example-project with a zonal MIG that is located in us-east1-b and has an autoscaler named example-autoscaler with an existing autoscaling policy.

PATCH https://compute.googleapis.com/compute/v1/projects/example-project/zones/us-east1-b/autoscalers?autoscaler=example-autoscaler

{
  "autoscalingPolicy": {
    "minNumReplicas": 0,
    "maxNumReplicas": 30,
    "scalingSchedules": {
      "workday-capacity": {
        "minRequiredReplicas": 10,
        "schedule": "30 8 * * Mon-Fri",
        "durationSec": 30600,
        "description": "Have at least 10 VMs every Monday through Friday from 8:30 AM to 5 PM UTC"
      }
    }
  }
}

To use this example for a regional MIG, make a similar request by using the regionAutoscalers.patch method and specify a region instead of a zone.

Onetime schedule

This example API call creates a schedule that runs once. This schedule specifies that you want a zonal MIG to contain at least 30 VM instances all day on January 30, 2030 for the America/New_York time zone. This API call assumes that you have an existing project named example-project with a zonal MIG that is located in us-east1-b and has an autoscaler named example-autoscaler with an existing autoscaling policy.

PATCH https://compute.googleapis.com/compute/v1/projects/example-project/zones/us-east1-b/autoscalers?autoscaler=example-autoscaler

{
  "autoscalingPolicy": {
    "scalingSchedules": {
      "example-onetime-schedule": {
        "minRequiredReplicas": 30,
        "schedule": "0 0 30 1 * 2030",
        "timeZone": "America/New_York",
        "durationSec": 86400,
        "description": "Schedule a minimum of 30 VMs all day for January 30, 2030"
      }
    }
  }
}

To use this example for a regional MIG, make a similar request by using the regionAutoscalers.patch method and specify a region instead of a zone.

Multiple schedules

This example API call uses the autoscalers.patch method for two schedules, workday-capacity and january-30-2030-schedule, in a single call. This API call assumes that you have an existing project named example-project with a zonal MIG that is located in us-east1-b and has an autoscaler named example-autoscaler with an existing autoscaling policy.

PATCH https://compute.googleapis.com/compute/v1/projects/example-project/zones/us-east1-b/autoscalers?autoscaler=example-autoscaler

{
  "autoscalingPolicy": {
    "scalingSchedules": {
      "workday-capacity": {
        "minRequiredReplicas": 10,
        "schedule": "30 8 * * Mon-Fri",
        "durationSec": 30600,
        "description": "Have at least 10 VMs every Monday through Friday from 8:30 AM to 5 PM UTC"
      },
      "january-30-2030-schedule": {
        "minRequiredReplicas": 30,
        "schedule": "0 0 30 1 * 2030",
        "timeZone": "America/New_York",
        "durationSec": 86400,
        "description": "Schedule a minimum of 30 VMs all day for January 30, 2030"
      }
    }
  }
}

To use this example for a regional MIG, make a similar request by using the regionAutoscalers.patch method and specify a region instead of a zone.

After a schedule is created, you might need to wait a few minutes to see its status information.

Listing your scaling schedules

You can view a list of the schedules for a MIG. The list shows each schedule's settings and status information.

Console

  1. In the Google Cloud console, go to the Instance groups page.

    Go to Instance groups

  2. Click the name of a MIG from the list.

  3. Click Edit.

  4. You can see the total number of schedules under Autoscaling schedules. Click Manage schedules to display the list of your existing scaling schedules.

When you are finished, you can close the list by clicking Done.

gcloud

To get details for a MIG, including the configurations and statuses of its scaling schedules, use the gcloud compute instance-groups managed describe command.

gcloud compute instance-groups managed describe MIG_NAME \
    [--zone=ZONE | --region=REGION]

Replace the following:

  • MIG_NAME: the name of a MIG with an existing autoscaling policy.
  • ZONE or REGION: Optional: the zone or region where your MIG is located.

Any scaling schedules for your MIG are listed in the output, as shown in the following example.

...
autoscalingPolicy:
  ...
  maxNumReplicas: 30
  minNumReplicas: 0
  ...
  scalingSchedules:
    example-onetime-schedule:
      description: Schedule a minimum of 30 VMs all day for January 30, 2030
      disabled: false
      durationSec: 86400
      minRequiredReplicas: 30
      schedule: 0 0 30 1 * 2030
      timeZone: America/New_York
    workday-capacity:
      description: Have at least 10 VMs every Monday through Friday from 8:30
        AM to 5 PM UTC
      disabled: false
      durationSec: 30600
      minRequiredReplicas: 10
      schedule: 30 8 * * Mon-Fri
      timeZone: UTC
...
scheduledScalingStatus:
    example-onetime-schedule:
        state: READY
        nextStartTime: '2030-01-30T00:00:00.000-05:00'
        lastStartTime: ''
    workday-capacity:
        state: READY
        nextStartTime: '2020-11-04T08:30:00.000-00:00'
        lastStartTime: '2020-11-03T08:30:00.000-00:00'
...

REST

To get details for an autoscaler, including the configurations and statuses of its scaling schedules, use the autoscalers.get method for a zonal MIG or the regionAutoscalers.get method for a regional MIG.

For example, make the following call to list the schedules for a zonal MIG that has an existing autoscaling policy:

GET https://compute.googleapis.com/compute/v1/projects/PROJECT/zones/ZONE/autoscalers?autoscaler=AUTOSCALER_NAME

Replace the following:

  • PROJECT: your project id.
  • ZONE: the zone where your MIG is located.
  • AUTOSCALER_NAME: the name of an existing autoscaler. This is often the same as the name of the MIG.

Any scaling schedules for your MIG are listed in the output, as shown in the following example.

{
  ...
  "autoscalingPolicy": {
    ...
    "minNumReplicas": 0,
    "maxNumReplicas": 30,
    ...
    "scalingSchedules": {
      "example-onetime-schedule": {
        "minRequiredReplicas": 30,
        "schedule": "0 0 30 1 * 2030",
        "timeZone": "America/New_York",
        "durationSec": 86400,
        "description": "Schedule a minimum of 30 VMs all day for January 30, 2030",
        "disabled": false
      },
      "workday-capacity": {
        "minRequiredReplicas": 10,
        "schedule": "30 8 * * Mon-Fri",
        "timeZone": "",
        "durationSec": 30600,
        "description": "Have at least 10 VMs every Monday through Friday from 8:30 AM to 5 PM UTC",
        "disabled": false
      },
      ...
    }
  },
  "scheduledScalingStatus": {
    "example-onetime-schedule": {
      "state": "READY",
      "nextStartTime": "2030-01-30T00:00:00.000-05:00",
      "lastStartTime": ''
    },
    "workday-capacity": {
      "state": "READY",
      "nextStartTime": "2020-11-04T08:30:00.000",
      "lastStartTime": "2020-11-03T08:30:00.000"
    },
    ...
  },
...
}

Editing a scaling schedule

You can edit an existing scaling schedule to change any of its settings except for the name of the schedule. You can also disable or re-enable a scaling schedule.

Console

  1. In the Google Cloud console, go to the Instance groups page.

    Go to Instance groups

  2. Click the name of a MIG from the list.

  3. Click Edit.

  4. You can see the total number of schedules under Autoscaling schedules. Click Manage schedules to display the list of your existing scaling schedules.

  5. Select the checkbox for the scaling schedule that you want to edit.

  6. At the top of the Scaling Schedules pane, click Edit.

  7. In the new Edit scaling schedule pane, modify the fields that you want to change. For more information about each field, see Creating a scaling schedule.

  8. When you are finished, click Save.

You might need to wait a few minutes before your changes are effective. You can monitor the status of your schedules by clicking Refresh at the top of the Scaling Schedules pane.

When you are finished, you can close the list by clicking Done.

gcloud

To update an existing scaling schedule, use the gcloud compute instance-groups managed update-autoscaling command. Use the --update-schedule flag to indicate the name of the schedule to update. Include the other flags as needed for your changes.

gcloud compute instance-groups managed update-autoscaling MIG_NAME \
    --update-schedule=SCHEDULE_NAME \
    [--schedule-cron="CRON_EXPRESSION"] \
    [--schedule-duration-sec=DURATION] \
    [--schedule-time-zone="TIME_ZONE"] \
    [--schedule-min-required-replicas=MIN_REQ_REPLICAS] \
    [--schedule-description="DESCRIPTION"] \
    [--zone=ZONE | --region=REGION]

Replace the following:

  • MIG_NAME: the name of a MIG with an existing scaling schedule.
  • SCHEDULE_NAME: the name of the existing scaling schedule.
  • CRON_EXPRESSION: Optional: the new start time and recurrence of this schedule formatted as a cron expression.
  • DURATION: Optional: the new duration, in seconds, that this schedule is active.
  • TIME_ZONE: Optional: the new IANA time zone for the schedule's start time—for example, Europe/Paris. The default value is UTC.
  • MIN_REQ_REPLICAS: Optional: the new number of minimum required instances that this schedule provides when it is active.
  • DESCRIPTION: Optional: a new description of the scaling schedule.
  • ZONE or REGION: Optional: the zone or region where your MIG is located.

REST

To update the existing scaling schedules for a MIG, use the autoscalers.patch method for a zonal MIG or the regionAutoscalers.patch method for a regional MIG.

For example, make the following call to edit a schedule for a zonal MIG that has an existing autoscaling policy:

PATCH https://compute.googleapis.com/compute/v1/projects/PROJECT/zones/ZONE/autoscalers?autoscaler=AUTOSCALER_NAME

{
  "autoscalingPolicy": {
    "scalingSchedules": {
      "SCHEDULE_NAME": {
        "minRequiredReplicas": MIN_REQ_REPLICAS,
        "schedule": "CRON_EXPRESSION",
        "timeZone": "TIME_ZONE",
        "durationSec": DURATION,
        "description": "DESCRIPTION"
      },
      ...
    }
  }
}

Replace the following:

  • PROJECT: your project id.
  • ZONE: the zone where your MIG is located.
  • AUTOSCALER_NAME: the name of an existing autoscaler. This is often the same as the name of the MIG.
  • SCHEDULE_NAME: the name of the existing scaling schedule.
  • MIN_REQ_REPLICAS: Optional: the new number of minimum required instances that this schedule provides when it is active.
  • CRON_EXPRESSION: Optional: the new start time and recurrence of this schedule formatted as a cron expression.
  • TIME_ZONE: Optional: the new IANA time zone for the schedule's start time—for example, Europe/Paris. The default value is UTC.
  • DURATION: Optional: the new duration, in seconds, that this schedule is active.
  • DESCRIPTION: Optional: a new description of the scaling schedule.

Disabling and re-enabling a scaling schedule

Scaling schedules are enabled by default. Disable a schedule if you want to prevent a schedule from being active but want to save its configuration. Re-enable a disabled schedule when you want to use it again.

If you do not need to store the schedule or have reached the 128 schedule limit for this MIG, delete the schedule. If you want to disable autoscaling for a MIG, turn off autoscaling.

Console

  1. In the Google Cloud console, go to the Instance groups page.

    Go to Instance groups

  2. Click the name of a MIG from the list.

  3. Click Edit.

  4. You can see the total number of schedules under Autoscaling schedules. Click Manage schedules to display the list of your existing scaling schedules.

  5. Select the checkboxes for the scaling schedules that you want to disable or enable.

  6. Disable or enable the selected schedules.

    • To disable the selected schedules:
      1. At the top of the Scaling Schedules pane, click Disable.
      2. In the new Disable schedules dialog, click Disable.
    • To enable the selected schedules:
      1. At the top of the Scaling Schedules pane, click Enable.
      2. In the new Enable schedules dialog, click Enable.

You might need to wait a few minutes before your changes are effective. You can monitor the status of your schedules by clicking Refresh at the top of the Scaling Schedules pane.

When you are finished, you can close the list by clicking Done.

gcloud

To disable or re-enable a scaling schedule, use the gcloud compute instance-groups managed update-autoscaling command.

Disabling a scaling schedule

To disable a scaling schedule, use the --disable-schedule flag.

gcloud compute instance-groups managed update-autoscaling MIG_NAME \
    --disable-schedule=SCHEDULE_NAME \
    [--zone=ZONE | --region=REGION]

Replace the following:

  • MIG_NAME: the name of a MIG with an existing scaling schedule.
  • SCHEDULE_NAME: the name of the scaling schedule that you want to disable.
  • ZONE or REGION: Optional: the zone or region where your MIG is located.

Re-enabling a scaling schedule

To re-enable a scaling schedule, use the --enable-schedule flag.

gcloud compute instance-groups managed update-autoscaling MIG_NAME \
    --enable-schedule=SCHEDULE_NAME \
    [--zone=ZONE | --region=REGION]

Replace the following:

  • MIG_NAME: the name of a MIG with an existing scaling schedule.
  • SCHEDULE_NAME: the name of the scaling schedule that you want to re-enable.
  • ZONE or REGION: Optional: the zone or region where your MIG is located.

REST

To disable or re-enable scaling schedules for a MIG, use the autoscalers.patch method for a zonal MIG or the regionAutoscalers.patch method for a regional MIG.

Disabling scaling schedules

For each scaling schedule that you want to disable, set the disabled field to true. For example, make the following call to disable a schedule for a zonal MIG that has an existing autoscaling policy:

PATCH https://compute.googleapis.com/compute/v1/projects/PROJECT/zones/ZONE/autoscalers?autoscaler=AUTOSCALER_NAME

{
  "autoscalingPolicy": {
    "scalingSchedules": {
      "SCHEDULE_NAME": {
        "disabled": true
      },
      ...
    }
  }
}

Replace the following:

  • PROJECT: your project id.
  • ZONE: the zone where your MIG is located.
  • AUTOSCALER_NAME: the name of an existing autoscaler. This is often the same as the name of the MIG.
  • SCHEDULE_NAME: the name of the scaling schedule that you want to disable.

Re-enabling scaling schedules

For each scaling schedule that you want to re-enable, set the disabled field to false. For example, make the following call to re-enable a schedule for a zonal MIG that has an existing autoscaling policy:

PATCH https://compute.googleapis.com/compute/v1/projects/PROJECT/zones/ZONE/autoscalers?autoscaler=AUTOSCALER_NAME

{
  "autoscalingPolicy": {
    "scalingSchedules": {
      "SCHEDULE_NAME": {
        "disabled": false
      },
      ...
    }
  }
}

Replace the following:

  • PROJECT: your project id.
  • ZONE: the zone where your MIG is located.
  • AUTOSCALER_NAME: the name of an existing autoscaler. This is often the same as the name of the MIG.
  • SCHEDULE_NAME: the name of the scaling schedule that you want to re-enable.

Deleting a scaling schedule

Remove scaling schedules from a MIG by deleting them. Delete schedules that you no longer need, such as schedules with an OBSOLETE status.

An autoscaling policy must always have at least one scaling signal. If no other autoscaling signal exists, then you cannot delete all scaling schedules without deleting the entire autoscaling policy. If you want to delete all scaling schedules but keep the autoscaling policy, you must add at least one scaling signal prior to deleting all the schedules.

If you want to prevent a schedule from becoming active and save it for future use, disable the schedule. If you want to delete the autoscaling configuration for a MIG, delete the autoscaler.

Console

  1. In the Google Cloud console, go to the Instance groups page.

    Go to Instance groups

  2. Click the name of a MIG from the list.

  3. Click Edit.

  4. You can see the total number of schedules under Autoscaling schedules. Click Manage schedules to display the list of your existing scaling schedules.

  5. Select the checkboxes for the scaling schedules that you want to delete.

  6. At the top of the Scaling Schedules pane, click Delete.

  7. In the new Delete schedules dialog, click Delete.

You might need to wait a few minutes before your changes are effective. You can monitor the status of your schedules by clicking Refresh at the top of the Scaling Schedules pane.

When you are finished, you can close the list by clicking Done.

gcloud

To delete a scaling schedule, use the gcloud compute instance-groups managed update-autoscaling command with the --remove-schedule flag.

gcloud compute instance-groups managed update-autoscaling MIG_NAME \
    --remove-schedule=SCHEDULE_NAME \
    [--zone=ZONE | --region=REGION]

Replace the following:

  • MIG_NAME: the name of a MIG with an existing scaling schedule.
  • SCHEDULE_NAME: the name of the scaling schedule that you want to delete.
  • ZONE or REGION: Optional: the zone or region where your MIG is located.

REST

To delete scaling schedules for a MIG, use the autoscalers.patch method for a zonal MIG or the regionAutoscalers.patch method for a regional MIG. For each scaling schedule that you want to delete, set the schedule configuration to null

For example, make the following call to delete a schedule for a zonal MIG that has an existing autoscaling policy:

PATCH https://compute.googleapis.com/compute/v1/projects/PROJECT/zones/ZONE/autoscalers?autoscaler=AUTOSCALER_NAME

{
  "autoscalingPolicy": {
    "scalingSchedules": {
      "SCHEDULE_NAME": null
      ...
    }
  }
}

Replace the following:

  • PROJECT: your project id.
  • ZONE: the zone where your MIG is located.
  • AUTOSCALER_NAME: the name of an existing autoscaler. This is often the same as the name of the MIG.
  • SCHEDULE_NAME: the name of the scaling schedule that you want to delete.

Concepts

This section describes concepts related to creating and using scaling schedules including schedule settings and details about how they work.

Scaling schedule settings

When you create and modify the scaling schedules for a MIG, each scaling schedule contains the following settings:

  • Minimum required instances: The number of VMs you require in the MIG when this scaling schedule is active.

    When this schedule is active, the MIG has at least this number of VMs, but might have more depending on other autoscaling settings. Additionally, the MIG size is always restricted by the minimum and maximum number of instances configured in the MIG's autoscaling settings.

  • Time zone: The location-based IANA time zone for interpreting the schedule's start time.

    A list of available values is defined by the IANA time zone database—for example, Europe/Paris. If no time zone is provided, UTC is used by default. Some IANA time zones observe daylight saving time. To learn how daylight saving time affects scaling schedules, see Daylight saving time.

  • Duration: The amount of time that you want the scaling schedule to be active.

    The scaling schedule is active from its start time for the configured duration. The minimum duration is 5 minutes. During this time, autoscaler scales the MIG to have at least as many VMs as defined by the schedule's required instances. After the specified duration, if the current capacity is no longer needed, the autoscaler starts removing VM instances after a default 10-minute stabilization period and following any configured scale-in controls.

    Scaling schedules are bidirectional. VMs are removed at the stop time that you specify and the group size is reduced to the minimum group size that you set, unless you have specified other autoscaling signals.

When using the Google Cloud CLI or REST, you must configure the start time and recurrence of a scaling schedule by using a cron expression. However, when using Google Cloud console, you can configure the start time and recurrence of a schedule either by using a cron expression or the following settings:

  • Start time: The time of day when you want the scaling schedule to become active.

    At the start time, the scaling schedule begins creating VMs whenever the MIG size is lower than the schedule's required instances. Set the start time to allow enough time for new VMs to boot and initialize. For example, if your workload takes 10 minutes from VM creation to begin serving, then set the start time to be 10 minutes earlier than the time you need the VMs to be ready.

  • Recurrence: If and how often the scaling schedule repeats, as defined by the following options:

    • Once: a onetime schedule that starts only on the selected date.
    • Every day: a repeating schedule that starts every day.
    • Every week: a repeating schedule that starts every week on the selected days of the week—for example, every Monday through Friday.
    • Every month: a repeating schedule that starts every month on the selected days of the month—for example, on the first day of every month. The last days of the month (days 29 through 31) are only effective during the months when those days occur.

Cron expressions

The start time and recurrence of a scaling schedule can be configured by using a cron expression. Cron is a time-based job scheduler originally used in Unix-like computer operating systems.

Syntax

A cron expression is a string comprising of five or six fields separated by whitespace. The following table defines the fields of a cron expression and the possible values for each field.

Minute Hour Day of the month Month Day of the week Year (optional)
0-59 0-23 1-31 1-12 (Jan-Dec)

where 1=Jan, 2=Feb, ... 12=Dec		 0-6 (Sun-Sat) or 1-7 (Mon-Sun)

where 0=Sun, 1=Mon, ... 6=Sat, 7=Sun 		2000-2036

or every year (*) if not provided

In addition to using these values, every field in a cron expression can also use special characters:

Special character Meaning Example
* any If the day of month field, day of the week field, and year field (if included) are each set to *, then the schedule starts everyday.
- range If the day of week field is set to 1-5 or Mon-Fri, then the schedule starts every week Monday to Friday.
, list If the month field is set to 5,7,9 or May,Jul,Sep, then the schedule starts every May, July, and September.
/ step If the month field is set to */3, then the schedule starts during the first month and every 3 months after that: January, April, July, and October.

When you write a cron expression, consider the following:

  • Whitespace is used to separate the fields of a cron expression. Remember to not add extra whitespace in a field that is using special characters.
  • When you specify both a day of the week and a day of the month (when neither field is set to *), the schedule uses the union of these values, not the intersection. For example, the schedule 0 8 1 * Mon starts at 8:00 AM on every Monday and on the first day of every month. That schedule does not start at 8:00 AM only on Mondays that are also the first day of the month.
  • You can use zero or more special characters in each field of a cron expression. For example, if you want a schedule to start every month except May, you can set the month field to 1-4,6-12 (a list of two ranges without whitespace).
  • When using ranges (-), you cannot use 0-7 or Sun-Sun for the day of the week field—to specify every day of the week, use *, 0-6, 1-7, Sun-Sat, or Mon-Sun instead.
  • When using steps (/), remember that cron expressions are stateless, so steps can sometimes be uneven. For example, if you want a schedule to run every 6 hours, you can set the hours field to */6, which behaves the same as listing every valid multiple of 6 (0,6,12,18). The steps are even because 24 hours is easily divided into 6-hour parts. However, if you set the hours field to */7, the schedule will not always start in 7-hour intervals. Instead, */7 behaves the same as listing every valid multiple of 7 (0,7,14,21), meaning that the schedule can start during the hours of 12 AM, 7 AM, 2 PM, and 9 PM.

Repeating schedules versus onetime schedules

Depending on how you write your cron expression, your schedule might run once or periodically repeat. To create a schedule that only runs once, you must specify the year field. For example, the cron expression 0 0 30 1 * 2030 configures a schedule to only start once, at midnight on January 30, 2030. To create a schedule that repeats, use special characters to describe when that schedule is to repeat. For example, the cron expression 30 8 * * Mon-Fri configures a schedule to start at 8:30 AM on every Monday, Tuesday, Wednesday, Thursday, and Friday.

Schedule status information

List your scaling schedules to view their status. A scaling schedule has the following status information:

  • Status (state): The current status of the scaling schedule, represented by one of the following states:
    • Ready (READY): This schedule is ready but not active yet. Autoscaler is awaiting the next occurrence of this schedule.
    • Active (ACTIVE): This schedule is currently active. Autoscaler is trying to provide this MIG with at least the required instances that this schedule specifies.
    • Obsolete (OBSOLETE): This schedule has expired; it will not be active again. Autoscaler is ignoring this schedule because there are no future occurrences.
    • Disabled (DISABLED): This schedule cannot be active because it was manually disabled. Autoscaler is ignoring this schedule until it is re-enabled.
  • Last occurrence (lastStartTime): The timestamp of the last time this scaling schedule became active. This is the actual date and time of the last activation. For example, if you have a disabled scaling schedule with a daily recurrence, 10 AM start time, and 2-hour duration, and you re-enable the schedule at 11 AM today, then the lastStartTime is 11 AM today.
  • Next occurrence (nextStartTime): The timestamp of the next time this schedule is set to become active.

Overlapping schedules

Depending on start time, recurrence, duration, and time zone, two or more scaling schedules might overlap. When more than one schedule is active, the autoscaler uses the greatest number of required VM instances from all active scaling schedules. This helps you ensure that there is always enough capacity to meet the needs of any active schedule. For example, if you have a schedule that recurs every day 8 AM to 10 AM requiring at least 10 VMs, but also have a one-off all-day schedule requiring at least 20 VMs, then autoscaler scales the MIG to at least 20 VMs on that day. This is useful when you have regular, recurring capacity requirements but sometimes need more to handle onetime events. This behavior lets you schedule capacity without worrying about it being overridden by other schedules with smaller VM requirements.

Additionally, an autoscaler always provides the greatest number of VM instances required by any active autoscaling signal. For more information, see Using an autoscaling policy with multiple signals.

Daylight saving time

The details of daylight saving time (DST) vary according to your scaling schedule's location-based IANA time zone. If your schedule's time zone observes DST, the schedule's start time automatically adjusts to stay synchronized with the time for the selected location. For example, if you create a scaling schedule that starts every day at 7 AM for the time zone America/New_York, then this schedule starts whenever it is 7 AM in New York, using Eastern Daylight Time or Eastern Standard Time accordingly.

Scaling schedules always run for the entire duration that you configure, but they might have slightly different start and end times on the days when DST observation switches. This happens because DST switches have transition periods—a timeframe that is skipped on the first day of DST and that occurs twice on the last day of DST. For example, if DST for your time zone begins at 2 AM and has a 1-hour transition, then the clock changes from 1:59:59 AM to 3 AM, so the timeframe 2 AM to 2:59:59 AM is skipped. If DST for that time zone ends at 2 AM, then the clock changes from 2:59:59 AM to 2 AM, so the timeframe 2 AM to 2:59:59 AM is repeated. The following statements assume that the time zone for your schedule uses a 1-hour DST transition.

On the first day of DST (when 1 hour is skipped), scaling schedules have the following changes:

  • Schedules that are active during the skipped hour still run for their entire duration, but to compensate for the skipped hour, these schedules end 1 hour later than their end time on days without DST transitions. Normally, the end time of a schedule is the sum of the schedule's start time and duration. But if DST skips 1 hour while a schedule is active, and the duration and start time remain constant, then the end time for that schedule is 1 hour more than the sum of the schedule's start time and duration.
  • Schedules that are configured to start during the skipped hour also start 1 hour later than their start time on days without DST transitions.

On the last day of DST (when 1 hour is repeated), scaling schedules have the following changes:

  • Schedules that are active during the repeat hour still run for their entire duration, but to compensate for the repeated hour, these schedules end 1 hour earlier than their end time on days without DST transitions. Normally, the end time of a schedule is the sum of the schedule's start time and duration. But if DST repeats 1 hour while a schedule is active, and the duration and start time remain constant, then the end time for that schedule is 1 hour less than the sum of the schedule's start time and duration.
  • Schedules that are configured to start during the transition period start at their configured time, but they do not restart during the second, repeated occurrence of the start time.

Most of the time zones that observe DST use a 1-hour transition period. But, if the time zone for your schedule uses a different transition period—such as the Australia/Lord_Howe time zone, which uses 30-minute transitions—you can expect the same behavior. Just replace 1 hour with the relevant DST transition time.

What's next