bq-cost-monitoring v1.1.5

Modified BigQuery Cost Monitor for Dataform

This is a modified version of the BigQuery Cost Monitor package that addresses issues with duplicate actions and canonical targets in Dataform projects. It also adds support for using existing datasets.

What's Changed

1. Added tracking of created tables/views: The package now keeps track of tables and views it has created to avoid duplicate actions.
2. Added support for existing datasets: A new useExistingDataset option allows you to use existing datasets instead of creating new ones.
3. Improved error handling: Better handling of schema and table name conflicts.
4. Configurable project database: A new projectDatabase option allows you to specify which GCP project to query for INFORMATION_SCHEMA data.
5. Optimized query performance: Added filters to reduce the amount of data processed in the query.
6. Fixed duplicate action name errors: Implemented a unique action name generation system to avoid conflicts.
7. Uses explicit table references: Removed all usage of ref() function in SQL queries for better compatibility.
8. Fixed template literal interpolation: Properly interpolates variables in SQL queries by pre-computing table names outside of template literals.

Installation

Since this is a modified version, you'll need to use it directly from the source files:

1. Copy the modified_index.js file to your project
2. Import it directly in your Dataform project

Usage with Existing Datasets

// Import the modified BigQueryCostMonitor class
const { BigQueryCostMonitor } = require("./path/to/modified_index");

// Initialize with an existing dataset
const costMonitor = new BigQueryCostMonitor({
  schema: "data_governance",  // Your existing dataset
  historyDays: 30,
  costPerTerabyte: 5.0,
  useExistingDataset: true,   // Enable existing dataset support
  projectDatabase: "my-gcp-project"  // Specify the project database (GCP project ID)
});

// Create tables with unique names
const mainTable = costMonitor.createCostMonitoringTable({
  name: "bq_cost_monitoring_table"  // Use a unique name
});

// Create views with unique names
costMonitor.createDailySummaryView({
  name: "bq_daily_cost_summary",
  sourceTable: mainTable.name
});

Configuration Options

The BigQueryCostMonitor constructor accepts the following options:

Troubleshooting

If you still encounter issues with duplicate actions or canonical targets:

1. Use unique table and view names: Make sure each table and view has a unique name across your entire Dataform project. Consider using a prefix like bq_ for all cost monitoring tables.

2. Check your Dataform configuration: Your Dataform project might have default schema settings that are causing conflicts. Check your dataform.json file for any default schema configurations.

3. Look for existing tables: Check if there are any existing tables in your BigQuery project with the same names as the ones you're trying to create.

4. Try a different schema: If possible, use a completely different schema name that doesn't exist in your project.

5. Debug with console logs: The modified package includes console logs that can help identify which tables are being created and which ones are being skipped due to duplicates.

Common Errors and Solutions

"Duplicate action name detected"

This error occurs when Dataform detects multiple actions (tables, views, etc.) with the same name. To fix:

• Use unique names for all tables and views
• Make sure you're not calling the same creation method multiple times
• Check if the table or view already exists in your Dataform project

This package now automatically generates unique action names when using existing datasets to avoid this error.

"Duplicate canonical target detected"

This error occurs when Dataform detects multiple objects targeting the same BigQuery table or view. To fix:

• Use the useExistingDataset: true option
• Use unique schema and table name combinations
• Check your Dataform project for any default schema settings

The latest version (1.1.2+) includes a fix that uses unique action names with the correct schema and table name configuration to avoid this error.

Performance Optimization

The package now includes filters to reduce the amount of data processed in the query:

• Excludes queries with zero bytes processed
• Excludes small queries (less than 1MB) that don't contribute significantly to costs

These optimizations can significantly reduce the amount of data processed (from 600GB+ to much less) while still providing accurate cost monitoring information.

Example Files

• existing_dataset_example.js: Shows how to use the modified package with existing datasets
• modified_index.js: The modified version of the package

Original Package

For reference, the original package documentation is available in the main README.md file.