Skip to main content

Set up PerfectScale for Commitments

This guide describes how to set up PerfectScale for Commitments for Amazon Web Services (AWS) and Google Cloud (GCP).

Prerequisites

  • Cloud provider is connected to DoiT Cloud Intelligence. For DoiT to detect AWS payer accounts and Google Cloud billing accounts with billing access and lists them during setup.

  • Your DoiT account has PerfectScale Commitments Admin permission. See Manage roles and users.

  • Your DoiT subscription includes the PerfectScale for Commitments entitlement, and you have an active PerfectScale for Commitments contract.

    PerfectScale for Commitments is part of Cloud Intelligence and is governed by the same Master License and Services Agreement; the contract is an addendum in which you agree to the savings-share fee. If your subscription includes the entitlement but you have no active contract, you'll see a message prompting you to contact your account team or DoiT support.

Step 1: Select cloud provider

Select your cloud provider and the commitment types you want to optimize.

  1. In the DoiT console, select Rate optimization > PerfectScale for Commitments, then select Connect.

  2. On the Cloud selection step, select all commitment types within your cloud providers that you want to add and optimize.

    Cloud providerAvailable commitment types
    AWSCompute Savings Plans, Database Savings Plans
    Google CloudCompute flexible CUDs

    Select commitment types

  3. Do one of the following:

    • If the cloud providers are successfully connected, select Next.

    • If the cloud provider connection fails, verify that you are using the correct DoiT customer and have granted the DoiT service identity the permissions described in Security and Data Policy.

Step 2: Select accounts

This step lists the accounts DoiT detected for the cloud providers you selected in Step 1: Select cloud provider. Only accounts that have not yet been onboarded are shown.

Note

If no accounts appear, either DoiT has not detected payer accounts (AWS) or billing accounts (Google Cloud) for your connected cloud provider, or all detected accounts are already onboarded.

When everything is already onboarded, a No accounts available message appears, and the Next button is disabled. See Prerequisites and Connect your cloud, or use Add new accounts in Account Settings.

  1. Select the accounts you want to use with PerfectScale for Commitments.

    Select payer accounts

    Note

    The AWS payer accounts must have PerfectScale for Commitments read permission to be selectable. See the next step for how to add the permission.

  2. If an orange warning icon appears next to an account, review and do the following for each cloud provider:

    • Read permission is missing: The payer account is missing the read permission and is disabled until you grant permissions. Select Configure and select the following checkboxes:

      • PerfectScale for Commitments - Read & Recommend (required): grants read-only permissions and enables visibility and recommendation features.

      • PerfectScale for Commitments - Purchases (optional): grants write permissions and enables auto-purchasing features.

      Then, select Update account. When read permission is verified, the account becomes selectable and shows a green check icon.

    • Linked accounts read permission is missing: The payer account is selectable, and you can configure the linked accounts' permissions at a later stage. Recommended: select Configure on each linked account that shows a warning, and enable PerfectScale for Commitments - Read & Recommend.

    For the full list of required IAM permissions, see Security and Data Policy.

    AWS missing read permission

  3. Select Next when the following are true for each cloud provider you enabled in Step 1:

    • AWS: At least one payer account is selected. Linked-account permission warnings do not block Next.

    • Google Cloud: At least one billing account is selected, and every selected billing account passes all health checks.

Resolve CUD export confirmation required

Use this procedure to enable Google Cloud CUD export when the health check shows CUD export confirmation required. PerfectScale reads spend-based CUD metadata from the BigQuery table Google creates when you enable CUD export in the Google Cloud console.

You need the Billing Account Administrator (roles/billing.admin) role on the Cloud Billing account.

  1. Open the Billing export page in the Google Cloud console.

  2. Select the Cloud Billing account you are onboarding for PerfectScale for Commitments.

  3. On the BigQuery export tab, find Committed use discounts (CUD) metadata and select Enable CUD export (or Edit settings if you are updating an existing export).

  4. Select the project to host the export, then select or create the BigQuery dataset.

  5. Select Save.

For full setup details and dataset location requirements, see Google Cloud Docs on Enable Cloud Billing export to the BigQuery dataset and Structure of CUD metadata export.

After you enable the export, return to the onboarding screen and select I've enabled CUD export. PerfectScale for Commitments cannot auto-detect whether CUD export is enabled.

Note

It may take a few hours before Google populates cud_subscriptions_export after you first enable CUD metadata export.

Resolve Billing export dataset access required

Use this procedure to connect the CUD export dataset when the health check shows Billing export dataset access required. After you enable Google Cloud CUD export, grant DoiT read access to the BigQuery dataset you have exported CUD metadata to. The dataset connection must be for a project linked to the billing account you are onboarding. DoiT matches Cloud Connect dataset connections to billing accounts by project membership.

  1. Select Configure on the Dataset connection check in this step, or open Integrations > Google Cloud > Manage > Google Cloud access & features.

  2. From the Connect drop-down, select Dataset.

  3. Select PS4Commitments CUD Export. You can expand the feature to view the required permissions.

  4. Select Generate gcloud commands.

  5. Enter the dataset name (the BigQuery dataset ID from your CUD metadata export) and a service account identifier following the naming conventions shown in the DoiT console.

  6. In Google Cloud Shell, select the project that hosts the CUD export dataset, then run the gcloud commands in sequence.

  7. Download doit_cmp_dataset_sa_details.json and select Upload file in the DoiT console.

  8. Wait until the connection shows a healthy PS4Commitments CUD Export status, then return to onboarding and confirm the Dataset connection check passes.

If you already have a dataset Cloud Connect connection for that project and dataset, select the menu next to it, select Edit, add PS4Commitments CUD Export, regenerate the gcloud commands, and update IAM permissions.

For the general dataset connection pattern, see Connect a dataset.

Step 3: Review and configure purchasing accounts

To enable purchases:

  1. Configure a Commitments purchasing account for each payer account. This is the AWS account where PerfectScale for Commitments buys Savings Plans. You can use the payer account itself.

    The purchasing account needs write permission (PerfectScale for Commitments - Purchases). If you see a permission warning, complete the permission steps in Step 2: Select accounts.

    Tip

    Prefer a purchasing account with little or no eligible compute usage. AWS applies Savings Plans to the purchasing account first, which can leave less discount for linked accounts.

    Note

    You can set up purchases later in Account Settings. To skip purchases for now, leave these settings empty and select Next.

  2. Select Next to complete the setup.

AWS and Google Cloud configure purchasing accounts

Once you have completed these steps, your account is onboarded, and PerfectScale for Commitments begins analyzing your usage to provide optimization recommendations.

  • On AWS, it may take up to 90 minutes before you see recommendations and other data in your dashboard.

  • On Google Cloud, it may take a few hours before you see recommendations and other data in your dashboard.