NAME
    gcloud alpha compute advice capacity - get capacity advice for Compute
        Engine resources

SYNOPSIS
    gcloud alpha compute advice capacity
        --provisioning-model=PROVISIONING_MODEL --size=SIZE
        --target-distribution-shape=TARGET_DISTRIBUTION_SHAPE
        (--instance-selection=INSTANCE_SELECTION
          --instance-selection-machine-types=[MACHINE_TYPE,...])
        [--accelerator-topology=ACCELERATOR_TOPOLOGY]
        [--max-run-duration=MAX_RUN_DURATION] [--region=REGION]
        [--zones=[ZONE,...]] [GCLOUD_WIDE_FLAG ...]

DESCRIPTION
    (ALPHA) Get capacity advice for Compute Engine resources.

    This command helps you view future resource availability for a specific
    number of VM instances, machine type, provisioning model, and zone. After
    you confirm resource availability, you can specify those configurations
    when you create VM instances. This action improves the success rate of your
    VM instance creation request.

EXAMPLES
    To check the availability of 100 n2-standard-32 Spot VMs in any single zone
    in the us-central1 region, run the following command:

        $ gcloud alpha compute advice capacity --region="us-central1" \
            --provisioning-model="SPOT" --size=100 \
            --instance-selection-machine-types="n2-standard-32" \
            --target-distribution-shape="any-single-zone"

    To check the availability of 50 Spot VMs, allowing either e2-standard-8 or
    e2-standard-16 machine types, distributed across us-central1-a and
    us-central1-b, run the following command:

        $ gcloud alpha compute advice capacity --region="us-central1" \
            --provisioning-model="SPOT" --size=50 \
            --instance-selection="name=my-selection,machine-type=e2-standard\
        -8,machine-type=e2-standard-16" --target-distribution-shape="any" \
            --zones="us-central1-a,us-central1-b"

    To check the availability of 10 ct5lp-hightpu-4t Flex-start VMs that will
    run for 1 day in any single zone in the us-central1 region, run the
    following command:

        $ gcloud alpha compute advice capacity --region="us-central1" \
            --provisioning-model="FLEX_START" --max-run-duration=1d \
            --size=10 \
            --instance-selection-machine-types="ct5lp-hightpu-4t" \
            --target-distribution-shape="any-single-zone" \
            --accelerator-topology="2x2"

REQUIRED FLAGS
     --provisioning-model=PROVISIONING_MODEL
        Specifies the provisioning model. PROVISIONING_MODEL must be one of:

         FLEX_START
            Compute Engine will attempt to allocate the requested capacity
            within a specific time frame. If it succeeds, the VMs will start
            running and keep running for a maximum of 7 days.
         SPOT
            Compute Engine may preempt a Spot VM whenever it needs capacity.
            Because Spot VMs don't have a guaranteed runtime, they come at a
            discounted price.

     --size=SIZE
        The total number of VMs being requested in the capacity query.

     --target-distribution-shape=TARGET_DISTRIBUTION_SHAPE
        Defines the distribution requirement for the requested VMs.
        TARGET_DISTRIBUTION_SHAPE must be one of:

         ANY
            When you specify ANY for VM instance creation across multiple
            zones, you specify that you want to create your VM instances in one
            or more zones based on resource availability.
         ANY_SINGLE_ZONE
            When you specify ANY_SINGLE_ZONE for VM instance creation, you
            specify that you want to create all VM instances in a single
            available zone. Compute Engine selects this zone based on resource
            availability and on any unused, matching reservation in your
            project. However, zonal resource availability constraints might
            prevent Compute Engine from creating all your requested VM
            instances. Use ANY_SINGLE_ZONE when the VM instances in your
            workloads need to frequently communicate among each other.
         BALANCED
            Prioritizes provisioning VMs in zones where resources are
            available. Compute Engine then distributes VMs as evenly as
            possible across zones. If resources are unavailable in one or more
            zones, then the actual distribution might be uneven. Use BALANCED
            to minimize the impact of zonal failures.

     Specifies the machine types for which advice is being sought.

     At least one of these must be specified:

       --instance-selection=INSTANCE_SELECTION
          Named selection of machine types. For example,
          --instance-selection="name=instance-selection-1,machine-type=e2-standard-8,machine-type=t2d-standard-8".

       --instance-selection-machine-types=[MACHINE_TYPE,...]
          Specifies a comma-separated list of preferred machine types for
          creating virtual machines.

OPTIONAL FLAGS
     --accelerator-topology=ACCELERATOR_TOPOLOGY
        Topology required to create a partition for VMs that have
        interconnected GPUs or TPUs. For example, 2x2 for ct5l-hightpu-4t.

     --max-run-duration=MAX_RUN_DURATION
        Maximum duration for which the requested VMs can run. This flag must be
        used when the requested provisioning model is FLEX_START.

        The duration is formatted as the number of days, hours, minutes, and
        seconds followed by d, h, m, and s respectively. For example, specify
        30m for a duration of 30 minutes or specify 1d2h3m4s for a duration of
        1 day, 2 hours, 3 minutes, and 4 seconds.

     --region=REGION
        Region of the resources to get advice on. If not specified, you might
        be prompted to select a region (interactive mode only).

        To avoid prompting when this flag is omitted, you can set the
        compute/region property:

            $ gcloud config set compute/region REGION

        A list of regions can be fetched by running:

            $ gcloud compute regions list

        To unset the property, run:

            $ gcloud config unset compute/region

        Alternatively, the region can be stored in the environment variable
        CLOUDSDK_COMPUTE_REGION.

     --zones=[ZONE,...]
        A comma-separated list of zones to query within the specified region,
        for example, us-central1-a,us-central1-b. If you omit this flag, then
        you view availability for your requested capacity across all zones in
        the region.

GCLOUD WIDE FLAGS
    These flags are available to all commands: --access-token-file, --account,
    --billing-project, --configuration, --flags-file, --flatten, --format,
    --help, --impersonate-service-account, --log-http, --project, --quiet,
    --trace-token, --user-output-enabled, --verbosity.

    Run $ gcloud help for details.

NOTES
    This command is currently in alpha and might change without notice. If this
    command fails with API permission errors despite specifying the correct
    project, you might be trying to access an API with an invitation-only early
    access allowlist.

