Provision the Data Warehouse and connect a BI tool
This walks through the full setup — provision the dataset, download credentials, and connect a BigQuery-compatible BI tool.
Before you start
- A user role with the Manage integrations permission
- A BI tool that supports Google BigQuery as a data source (Looker Studio, Tableau, Metabase, Sigma, etc.)
- Somewhere safe to store a service account JSON key (a password manager or your tool’s credentials store)
Step 1 — Provision the warehouse
- Open Settings → Data Warehouse.
- Click Provision Data Warehouse.
Atender creates a dedicated BigQuery dataset for your tenant, generates the SQL views, and creates a service account. You’ll see a “Data warehouse provisioned” confirmation with the number of views created.
The page then shows the Connection Details card with three values:
- Project — the Google Cloud project ID
- Dataset — your tenant’s dataset name
- Service Account — the service account email Atender created
Each value has a copy button next to it.
Step 2 — Generate credentials
- Scroll to the Credentials card.
- Click Download New Key.
Your browser downloads a JSON file named something like bigquery-credentials-<dataset>.json. Save it somewhere secure — Atender doesn’t keep a copy you can re-download.
The key count on the credentials card increments by one. Generate additional keys if you have multiple BI tools or environments.
Step 3 — Connect a BI tool
The exact steps depend on the tool, but the flow is always the same: add a BigQuery data source, point it at your project and dataset, and authenticate with the JSON key.
In your BI tool’s data source / connection setup:
- Choose Google BigQuery as the data source type.
- Authenticate with the service account JSON key file you just downloaded.
- Enter the Project ID from the Connection Details card.
- Enter the Dataset name from the Connection Details card.
- The tool should list the available views — pick the ones you want to use.
The tool can now query Atender data with SQL.
For column-level documentation — what each view contains, how columns are typed — the How to connect card has a link to the Atender Analytics docs at https://atender-labs.gitbook.io/atender-labs-docs/.
Verify it worked
In your BI tool, run a sanity-check query against a view — for example, count conversations created this week. The number should match what you see in the Atender Analytics module for the same window. If it doesn’t, see Troubleshooting below.
Manage credentials over time
- Add a key — click Download New Key again. Atender stores no limit on the number of active keys; each one works independently.
- Revoke all keys — click Revoke All Keys under the Credentials card and confirm in the dialog. This invalidates every active service account key. BI tools connected with revoked keys will start returning auth errors. Use this if a key has leaked.
- Rotate — generate a new key, switch your BI tool to use it, then revoke and replace.
Views update automatically
When Atender ships new views or updates existing ones, the current views reach your dataset automatically within about an hour of a deploy. You don’t need to do anything, and running queries and credentials aren’t affected.
You only need to click Re-provision Views in the Connection Details card if:
- You want the latest views right away instead of waiting for the automatic update.
Re-provisioning updates view definitions in place without touching credentials or breaking running queries.
If the workspace was renamed after you provisioned the warehouse, the automatic update skips it and the Connection Details card shows Update available. In that case Re-provision Views creates a new dataset under the new name instead of updating the old one. Your service account keys keep working, but update the Dataset in your BI tool afterwards.
Troubleshooting
-
Symptom: Provisioning fails with “Provisioning failed.” Fix: Wait a minute and try again — provisioning calls Google Cloud APIs and can briefly fail under load. If it keeps failing, contact Atender support.
-
Symptom: BI tool authentication fails with “permission denied.” Fix: The service account is scoped only to your tenant’s dataset, not the entire project. Make sure the BI tool is querying tables prefixed with your dataset name. Wildcard queries across other datasets will fail.
-
Symptom: Numbers in the BI tool don’t match Atender’s Analytics module. Fix: Check the time-window definitions and timezone. Atender Analytics uses the timezone of the person viewing Analytics, based on their browser/account context. BigQuery timestamps remain UTC by default, so convert in the query or adjust the BI tool’s timezone setting to the same IANA timezone used by the report viewer, such as
Europe/Oslo, when reconciling hourly or daily buckets. Different viewers in different timezones may see different day or hour groupings for the same underlying UTC events. -
Symptom: Your dataset still shows old behavior more than an hour after a deploy — for example, deleted contacts are still visible. Fix: Click Re-provision Views in the Connection Details card to update the views manually. If the old behavior persists, contact Atender support.
-
Symptom: “Key generation failed.” Fix: Usually a transient Google Cloud error. Try again. If it persists, contact support.
-
Symptom: Want to undo provisioning entirely. Fix: There’s no self-service de-provision. Contact support to tear down the dataset.