Database Metrics

This page documents Prometheus metrics related to user databases in the Yellowbrick system. These metrics reflect internal catalog states, transaction aging, and user data sizing across distributed storage.

Purpose

These metrics are primarily used to:

• Monitor database size and catalog footprint
• Detect potential transaction ID wraparound risks
• Track compression and table count trends across databases

They are valuable for capacity planning, transaction maintenance (e.g. VACUUM), and debugging database-level anomalies or storage bloat.

Metrics

Name	Type	Freq	Labels	Description
yb_database_age_frozen_transaction_id	gauge	1h	database_oid	Age of the oldest transaction ID that has not been frozen in the database
yb_database_age_minimum_multi_transaction_id	gauge	1h	database_oid	Age of the oldest multi-transaction ID that has been replaced with a transaction ID in the database
yb_database_catalog_data_bytes	gauge	1h	database_oid	Catalog and metadata size of the database
yb_database_table_count	gauge	1h	database_oid	Number of user tables in each database
yb_database_user_data_compressed_bytes	gauge	1h	database_oid	Compressed user data size for the database (worker nodes)