Airflow DagRuns

This page provides drill-down visibility into DAG execution history.

Pulse enables full-stack observability for Airflow on Kubernetes by combining:

• DAG-level monitoring
• Task-level drilldowns
• Scheduler and executor health tracking
• Resource utilization visibility
• Performance trend analysis
• Operator-level metrics

This allows faster root cause analysis, improved SLA tracking, and proactive workload optimization.

Steps

1. In the Pulse UI, go to Airflow on the left pane.
2. Select DagRuns.

DAGs

• DAG ID — Unique DAG name
• Owner — DAG owner
• Recent Runs — Status indicators of the latest runs
• Schedule — Cron or preset schedule
• Paused — Whether DAG is paused
• Start Date — Initial activation date

Runs Section

For each DAG run:

• Run ID — Unique run identifier
• State — Success, Failed, Running, etc.
• Run Type — Scheduled, Manual, Backfill
• Started At — Execution start time
• Queued At — Time entered queue
• Ended At — Completion time

DAG Run Details

Provides deep observability for a specific DAG run.

Run Summary

Run Information

• Run ID
• Run Type
• State

Timing Details

• Started At
• Queued At
• Ended At
• Duration

DAG Run Trends

Compares:

• Elapsed Time
• Queued Duration

Across historical runs.

Tasks Table

For each task:

Task Information

• Task ID
• Job ID
• Operator

Execution Status

• Previous State
• Current State

Timing Details

• Queue
• Started At
• Ended At
• Duration

Resource and Scheduling

• Pool
• Pool Slots
• Priority Weight

Task Tries

• Shows retry attempts and retry behavior trends.

Landing Times

• Displays task scheduling, landing times, and execution distribution.

Execution Timeline

Visual timeline view showing:

• Task execution order
• Parallelism
• Dependencies
• Branch behavior