Skip to main content

Analyze BigQuery activities using audit data

Overview​

Google Cloud generates BigQuery audit logs to record relevant administrative activities and accesses. These audit logs provide insight into operational concerns related to your BigQuery usage such as creating or deleting a table, purchasing slots, or running a load job.

The DoiT BigQuery Intelligence feature sets up a BigQuery audit log sink for widgets displayed on the BigQuery Intelligence dashboard that simplifies the analysis of BigQuery usage. You can also leverage the audit logs collected by the BigQuery Intelligence to perform your own analysis in Cloud Analytics reports, using BigQuery Intelligence as the data source.

Required permission​

  • Cloud Analytics User

Build reports using BigQuery Intelligence data​

To build a report with data from the BigQuery Intelligence audit log sink:

  1. Make sure you have already set up BigQuery Intelligence.

  2. Open a report. Select Change at the top of the left-hand sidebar.

    Update data source

  3. Select BigQuery Intelligence as the data source.

  4. You can choose to Create a new report or Update the current one. Note that updating the report will lose all its current configurations.

Metrics and dimensions​

Cloud Analytics reports use different sets of metrics and dimensions for billing data and BigQuery audit logs.

Metrics​

The basic metrics and extended metrics for BigQuery audit logs utilize fields of the jobStatistics object in BigQuery AuditData.

Basic metrics​

  • Cost: Processed bytes (totalBilledBytes) converted to TiB multiplied by the on-demand scan price.

  • Usage: Processed bytes (totalBilledBytes) converted to TiB, adjusted by the job's CPU usage.

Extended metrics​

  • Slots Used: The total number of slot-ms consumed divided by the query duration: totalSlotMs Γ· (endTime βˆ’ startTime).

  • Total Slots Ms: The total number of slot-ms consumed by the query job. It maps to totalSlotMs.

  • Total Load Output Bytes: Total bytes loaded for an import job. It maps to totalLoadOutputBytes.

  • Total Tables Processed: Total number of unique tables referenced in the query. It maps to totalTablesProcessed.

  • Total Billed Bytes: Processed bytes, adjusted by the job's CPU usage. It maps to totalBilledBytes.

  • Total Processed Bytes: Total bytes processed for a job. It maps to totalProcessedBytes.

Dimensions​

Below are the available standard dimensions when the report's data source is BigQuery Intelligence:

  • Event name: Name of the event. It maps to jobCompletedEvent.eventName in BigQuery AuditData.

  • Job Status: State of a job: PENDING, RUNNING, or DONE. It maps to jobStatus.state in BigQuery AuditData.

  • Project/Account name: Human-readable Google Cloud project name of the completed job. It maps to jobName.projectId in BigQuery AuditData.

  • Query Priority: Priority given to the query: QUERY_INTERACTIVE or QUERY_BATCH. It maps to jobConfiguration.query.queryPriority in BigQuery AuditData.

  • Region: Location of the completed job. It maps to jobName.location in BigQuery AuditData.

  • Reservation: Reservation name or unreserved for on-demand resource usage. It maps to JobStatistics.reservation in BigQuery AuditData.

  • Resource: URI for the referenced resource. For example, a table created by using an insert job reports the resource URI of the table. It maps to protoPayload.resourceName in BigQueryAuditMetadata messages.

  • Statement Type: Type of the statement. Example: SELECT, INSERT, CREATE_TABLE, CREATE_MODEL. It maps to jobConfiguration.query.statementType in BigQuery AuditData.

  • Caller IP: IP address of the caller. It maps to requestMetadata.callerIp in audit logs.

  • User: The email address of the authenticated user (or service account on behalf of third party principal) making the request. It maps to authenticationInfo.principalEmail in audit logs.

Examples​

Break down usage by user​

Usage-by-user

Identify hourly heavy queries by job ID​

Usage-by-user

See also​

Last updated on