Google Vertex AI
Connect Google Vertex AI usage, project, and region metrics to Doow.
Connect Google Vertex AI when Doow needs usage, cost, project, SKU, and region signals from Google Cloud billing data.
What you need
- A Google Cloud billing account that pays for the Vertex AI usage you want Doow to read.
- Cloud Billing export to BigQuery enabled for that billing account.
- A service account with read access to the billing export dataset.
- A project where the service account can run BigQuery jobs.
- Optional: one or more consumer project IDs to scope reporting to specific Vertex AI projects.
Use billing export for Google Cloud usage
Google Cloud exports detailed billing data to BigQuery with fields such as services, SKUs, projects, locations, usage, cost, and currency. Doow reads that export for Vertex AI reporting; it does not need access to model request payloads.
Create a read-only service account
- Go to the Google Cloud Console.
- Open IAM & Admin, then Service Accounts, then Create Service Account.
- Name it something identifiable, such as
doow-vertex-readonly. - Grant the service account read access to the BigQuery dataset that receives Cloud Billing export data.
- Grant the service account permission to run BigQuery jobs in the billing project.
- If your approval process requires billing-account visibility, grant Billing Account Viewer at the billing account level.
- Open Keys, then Add key, then Create new key.
- Choose JSON and download the key file.
When connecting in Doow, you will paste the contents of this JSON key file.
What Doow reads
| Field | Description |
|---|---|
| Usage date | The day the usage occurred |
| Service and SKU | Which Vertex AI service or SKU generated usage |
| Usage amount | Usage quantity in the unit reported by Google Cloud Billing |
| Cost | Cost estimate and currency from the billing export |
| Consumer project | The Google Cloud project that generated the usage |
| Location | Region or location fields from the billing export |
| Labels and credits | Labels, credits, or adjustments where present in the export |
Doow reads Vertex AI usage from the Cloud Billing export dataset and aggregates it at the day level by billing account, project, service, SKU, and location where those fields are available.
What Doow does not read
- The content of any model requests or responses
- Cloud Storage objects, databases, model artifacts, or unrelated Google Cloud data
- Payment methods, invoices, or billing contact information
- Vertex AI endpoint configuration or model settings
How to connect
- Go to Company Settings, then Integrations in your Doow workspace.
- Find Google Vertex AI and select Connect.
- Enter the billing project ID that contains the BigQuery billing export.
- Enter the billing export dataset and table details if prompted.
- Paste the contents of the service account JSON key.
- Optional: enter consumer project IDs to limit reporting to specific Vertex AI projects.
- Select Connect.
- Wait for Doow to verify the credentials and begin the initial usage sync.
Multi-project setups
If Vertex AI usage is spread across multiple Google Cloud projects under the same billing account, enter those project IDs in the Consumer project IDs field. Doow reads usage for those projects from the same billing export and shows per-project breakdowns.
Confirm the sync worked
After connecting, open the integration detail page from Company Settings, then Integrations. A healthy cloud sync shows a connected state, a recent sync timestamp, and usage rows for the expected billing account, project, service, SKU, or location.
If the connection succeeds but no usage appears, verify that Cloud Billing export is enabled, the export table contains Vertex AI rows, and the service account can read the BigQuery dataset.
Troubleshooting
Use these checks when cloud usage data is missing or incomplete:
- Confirm the credential belongs to the intended cloud account, subscription, or project.
- Confirm it can read the required monitoring, billing, or usage export data.
- Confirm the expected regions, deployments, or models have recent usage.
- Check whether the cloud provider reports usage with a delay.
- Check the integration event log for permission or sync errors.
Google Vertex AI-specific checks
| Symptom | Likely cause | Next action |
|---|---|---|
| No project usage appears | The billing export does not include the project or the project filter is too narrow | Confirm the consumer project IDs and billing account |
| Some usage is missing | Usage exists in a different billing account or export table | Connect the billing account or export that owns that usage |
| Billing data is unavailable | The service account cannot read the BigQuery export dataset | Grant dataset read access and BigQuery job permission before reconnecting |
| Recent usage is missing | Cloud Billing export has not delivered the latest rows yet | Check an earlier date range before rotating credentials |
Disconnecting
Go to Company Settings, then Integrations, find Google Vertex AI, and select Disconnect. Doow deletes the stored credentials immediately. Delete the service account key or remove its BigQuery and billing roles in Google Cloud when you want to remove provider-side access too.
Next steps
After cloud usage appears, connect an identity provider if Doow needs to match Vertex AI usage to users and app ownership. Connect HRIS when the review needs department or employment context.