Monitor pipelines

This section describes monitoring and observability features for Lakeflow Spark Declarative Pipelines.

Topic	Description
Monitor using the UI	Observe the progress and status of pipeline updates, and alert on the success or failure. View metrics for streaming sources, like Apache Kafka and Auto Loader.
Event log	Extract detailed information on pipeline updates such as data lineage, data quality metrics, and resource usage using the pipeline event log. Additionally, see the schema for the event log.
Query history	Inspect and diagnose query performance by looking at the query history.
Custom monitoring	Define custom actions to take when specific events occur using event hooks.

Additionally, there are troubleshooting topics for specific scenarios.

Topic	Description
Recover a pipeline from streaming checkpoint failure	Recover a pipeline that has an invalid or corrupted streaming checkpoint.
Fixing high initialization times in pipelines	Fix high initialization times for a pipeline by splitting and load balancing flows across pipelines.