Troubleshoot the Google Ads connector
This feature is in Beta.
Learn how to troubleshoot common errors with the managed Google Ads connector in Lakeflow Connect.
I've authenticated, but I can't see my Google Ads data
Cause: You might be authenticating to the wrong Google account. Alternatively, the account might not have access to Google Ads.
Solution:
- Confirm that you are authenticating to the correct Google account. You can edit the connection and reauthenticate with the correct account, or recreate the connection to reassess the OAuth flow.
- Confirm that the account has Google Ads access. Log into the Google account and verify that it has permissions to access your Google Ads data.
Error: GOOGLE_ADS_ACTION_NOT_PERMITTED
Error message:
The authenticated Google Ads user or the Google Cloud project that owns the OAuth client is not permitted to perform this action on the requested resource or to call this method (authorization_error: ACTION_NOT_PERMITTED or CLOUD_PROJECT_NOT_APPROVED_FOR_PRODUCTION).
Cause: Either the authenticated user doesn't have access to the account, or the Google Cloud project that owns your OAuth client has only Test access to the Google Ads API. A project with Test access can't read production Google Ads accounts.
Solution:
- Confirm that the authenticated user has access to the Google Ads account.
- Check the access level of the Google Cloud project that owns your OAuth client. See Confirm the Google Ads API access level. If it shows Test, apply for Basic or Standard access.
Redirect URI error during OAuth setup
Error message:
Error: redirect_uri_mismatch
Cause: The redirect URI configured in your Google Cloud Console OAuth client doesn't match the Databricks workspace URL.
Solution:
- Visit the Google Cloud Console at https://console.cloud.google.com.
- Navigate to APIs & Services /> Credentials.
- Click your OAuth client credential.
- Under Authorized redirect URIs, confirm that it contains the following pattern:
https://<workspace-url>/login/oauth/google.html. Replace<workspace-url>with your Databricks workspace URL (for example,https://dbc-a1b2c3d4-e5f6.cloud.databricks.com). - Save the changes and try authenticating again.
Error: SAAS_CONNECTION_ERROR.INVALID_CREDENTIALS
Error message:
Failed to make a connection to the Google Ads source. Error code: INVALID_OAUTH_CREDENTIALS. Invalid credentials. Ensure the source connection credentials are provided correctly.
Cause: Google Ads rejected the connection's OAuth credentials. For example, the user revoked the app's access, the user's password changed, or the OAuth client was deleted.
Solution: Edit the connection and reauthenticate. If you use your own OAuth client, confirm that it still exists in the Google Cloud console.
Error: GOOGLE_ADS_USER_PERMISSION_DENIED
Error message:
The authenticated Google Ads user does not have permission to access the requested customer account (authorization_error: USER_PERMISSION_DENIED). Verify the authenticated user is linked to the account with at least read access.
Cause: The authenticated user can't access the customer account that the pipeline reads. A common cause is that manager_account_id isn't set, or isn't the manager account that the customer account is linked to. This also happens when the account was unlinked from the manager account that the pipeline uses, was canceled, or was moved to another manager account, or when the user's access was removed.
Solution:
- In Google Ads, confirm that the customer account is linked to the manager account that the pipeline uses, and that the authenticated user has at least read access to it.
- If the account is no longer available, remove its tables from the pipeline. Otherwise, reauthenticate the connection with a user who has access.
Error: GOOGLE_ADS_QUOTA_EXCEEDED
Error message:
The Google Ads API quota has been exhausted (rate name: <rateName>, scope: <rateScope>). ...
Cause: The pipeline made more Google Ads API requests than the quota allows. Google Ads usually asks the client to wait longer than the pipeline's retry window, so the update fails instead of retrying.
Solution:
- Reduce the number of tables that the pipeline ingests, or schedule the pipeline less often.
- Wait before you rerun the pipeline. If the error message reports a retry delay, wait at least that long. A delay of
0means that Google Ads didn't report one. - If you use your own developer token and often reach the quota, apply for a higher access level for the token. See API limits and quotas in the Google Ads API documentation.
Error: GOOGLE_ADS_INVALID_CUSTOM_REPORT_CONFIGURATION
Error message:
Invalid custom_report configuration. <reason> ...
Cause: The custom_report_options of a custom report table can't be queried. The reason names the problem. Common reasons are:
- A field name is invalid, or the field can't be selected from the chosen
resource. - No timeline segment is selected. Each custom report needs
segments.date,segments.week, orsegments.month. - A field is selected from an attributed resource without that resource's
resource_name. - The
sync_start_dateis more than 37 months ago for a report segmented bysegments.dateorsegments.week.
Solution: Fix the option that the reason names. To test the configuration without moving data, run databricks pipelines start-update <pipeline-id> --validate-only. See databricks pipelines start-update. For the option rules, see Custom report configuration options. To report on data older than 37 months, use segments.month.
Error: GOOGLE_ADS_API_ERROR
Error message:
The Google Ads API returned an error (<googleAdsErrorCode>): <googleAdsErrorMessage> ...
Cause: Google Ads rejected a request with an error that the connector doesn't classify further. The error code and message usually come from Google Ads. One code comes from the connector instead: UNRECOGNIZED_ACCOUNT_TIME_ZONE means that Google Ads returned no time zone for the account, or one that the connector doesn't recognize, so the connector can't compute report dates in the account's time zone.
Solution: Look up the error code on the Common errors page of the Google Ads API documentation, and adjust the connection or the affected table. For UNRECOGNIZED_ACCOUNT_TIME_ZONE, which isn't listed there, or if the error persists, contact Databricks support.
Error: MANAGED_INGESTION_PIPELINE_SPEC_VALIDATION_ERROR.OBJECT_SETTING_NOT_SUPPORTED
Error message:
These settings are not supported on these <sourceType> objects: <violations>. Remove them from these objects. A table_configuration setting on a table can also come from the pipeline's table_configuration.
Cause: The pipeline sets an option on a table, schema, or catalog that doesn't support it. For example, a row_filter, primary_keys, or sequence_by on a report table, custom_report_options on a table that isn't a custom report, or sync_start_date or lookback_window_days on a resource or event table. The error lists each object with the settings to remove, for example ad: lookback_window_days, sync_start_date; campaign_report: row_filter.
Solution: Remove the listed settings from those objects. If a table doesn't set row_filter, primary_keys, or sequence_by itself, check the table_configuration of the pipeline, which the table inherits when it has none of its own. lookback_window_days and sync_start_date come from the table, from its schema, or otherwise from the first table or schema in the pipeline specification that sets Google Ads connector options. For the rules, see Table configuration limitations.
Error: MANAGED_INGESTION_CONNECTOR_OPTIONS_CHANGED_REQUIRE_FULL_REFRESH
Error message:
The <options> connector option(s) of the <sourceType> destination table `<tableName>` cannot be changed during an incremental update. Perform a full refresh of the table, restore the original options, or create a new destination table.
Cause: You changed a custom report's resource, resource_fields, segments, or metrics, or a report table's sync_start_date. These options define the rows that the table holds, so an incremental update can't apply the change. The error names the changed options. A sync_start_date that the table inherits from its schema, or from the first table or schema in the pipeline specification that sets Google Ads connector options, counts too. Changing that value can fail every report table that inherits it.
Some custom report changes fail earlier with SAAS_INCOMPATIBLE_SCHEMA_CHANGES_DURING_INIT instead, for example removing a metric or adding a segment. They need a full refresh too.
Solution: Do one of the following:
- Run a full refresh of only this table to re-ingest it with the new options. To refresh selected tables, see Fully refresh target tables. A full refresh removes history that Google Ads no longer serves, so copy that history to another table first. See Older history is missing after a full refresh.
- Restore the original options.
- Ingest the new configuration into a new destination table.
Error: CANNOT_FILTER_OUT_REQUIRED_COLUMN or CANNOT_ADD_COLUMN_TO_INCLUDE_COLUMNS
Error message:
Pipeline definition does not include required columns (<columnNames>) in the table <qualifiedTableName> of source <sourceType> for ingestion. Please add them to include_columns or remove them from exclude_columns.
New columns (<columnNames>) are added to include_columns field in pipeline definition for the table <qualifiedTableName>. Please full refresh the table to avoid potential data loss or remove these columns from include_columns.
Cause: The column selection removes a primary key or cursor column that the table needs, or adds columns to include_columns on a table that already exists. Earlier rows have no values for added columns.
Solution: Keep the primary key and cursor columns in the selection. To add columns to an existing table, run a full refresh of only that table, or remove the new columns from include_columns. To refresh selected tables, see Fully refresh target tables. A full refresh removes history that Google Ads no longer serves, so copy that history to another table first. See Older history is missing after a full refresh.
A report table has values that no longer match Google Ads
Cause: Google Ads changed data that is older than the lookback window, or restated a row to all-zero metrics. The connector re-reads only the lookback window and does not delete report rows, so the table keeps the earlier values. See Late data changes and deleted rows limitations.
Solution:
- To refresh the affected rows one time, run a full refresh of only that table. To refresh selected tables, see Fully refresh target tables. A full refresh removes history that Google Ads no longer serves, so copy that history to another table first. See Older history is missing after a full refresh.
- If changes often arrive later than the lookback window, increase
lookback_window_daysto match your conversion window.
Older history is missing after a full refresh
Cause: A full refresh ingests only what Google Ads still serves. Google Ads keeps hourly, daily, and weekly report data for 37 months, click data for 90 days, and lead form submissions for 60 days. History older than that is removed from the table. See Data retention limitations.
Solution:
- If you kept a copy of the table from before the full refresh, restore the older history from that copy.
- If you don't have a copy, contact Databricks support to assess recovery options.
- Before future full refreshes, copy the history that Google Ads no longer serves to another table.
- For a custom report that needs data older than 37 months, use
segments.month. This returns monthly totals for those periods. It doesn't restore the original hourly, daily, or weekly rows.
click_report is missing clicks after the pipeline was paused
Cause: Google Ads serves click data for only the last 90 days, and the connector reads at most the last 88 days. If the pipeline doesn't run successfully for more than 88 days, clicks between the last successful update and 88 days ago can't be ingested.
Solution: The missed dates can't be recovered. Run the pipeline on a schedule, and fix a failed update before 88 days pass after the last successful update.