Model monitor commands¶

A model monitor monitors the performance of machine learning models you have deployed in Snowflake. Model behavior can change over time due to input drift, stale training assumptions, and data pipeline issues, as well as changes to the underlying hardware and software and the fluid nature of traffic. Model monitors allow you to track the quality of your models across multiple dimensions, such as performance, drift, and volume.

Snowflake supports two types of model monitors:

• Model version monitor — monitors one model version using a source table or view. See ML Observability.
• Gateway model monitor — monitors inference services behind a Snowflake Gateway using auto-captured inference logs. See Gateway Monitoring & A/B Testing.

For details on model monitor commands, see the following:

• CREATE MODEL MONITOR
• DROP MODEL MONITOR
• DESCRIBE MODEL MONITOR
• ALTER MODEL MONITOR
• SHOW MODEL MONITORS