Reporting Airflow doesn't ship with — but your CIO expects

What enterprise schedulers gave you, and Airflow doesn't

Control-M BHOM (Helix Operations Management) gives you SLA dashboards, batch-financial reports, capacity heatmaps, and root-cause links for incident triage. Airflow's UI is excellent for individual DAG debugging but does not cover the Monday-morning ops review use case out of the box. The gap is real, and it surprises teams that migrated without planning for it.

The BatchFoundry reporting stack

Apache Superset (or Metabase, your choice) as the BI front end. A read-replica of the Airflow metadata database so reporting load doesn't affect scheduling. A curated set of dimensional views over dag_run, task_instance, sla_miss, xcom, and log tables.

Standard dashboards we ship

• SLA scorecard by DAG — breaches, near-misses, trend over 30 days
• Top-N flaky tasks by failure rate — with drill-down to log samples
• Queue depth by hour — identify persistent resource bottlenecks
• Batch-financials estimator — compute hours × hourly cost per worker pool

What BatchFoundry delivers

• Reporting database setup with proper indexes and retention policies
• Pre-built dashboards as code (Superset YAML or Metabase JSON — your choice)
• Training for your operations team on the new reporting stack
• Optional: a custom Sunrise Report emailed every weekday morning with the prior 24-hour batch summary

Pitfalls we plan around

• Querying the live Airflow metadata DB at scale — use a read-replica
• Time-zone normalization across multi-region Airflow clusters
• Retention policy on task_instance (default: infinite — bad at scale)
• PII / PHI in XCom payloads — mask before surfacing in reports