DE//AGENT
Explore
Tutorial Lesson 10 of 10 · The Orchestration Series community intermediate

Capstone — the checklist that decides whether a DAG ships

Close the series by running the full review pass over your own pipeline: the idempotency test per task, partitioned or MERGE writes, deliberate catchup, backoff retries, reschedule sensors, bounded concurrency — and a documented runbook for the day history needs fixing.

At a glance

When Signing off a DAG before it merges or takes production traffic; Auditing an inherited DAG for the first time
Where Airflow / Orchestration
Why Ten lessons reduce to one list — if the checklist passes, the DAG survives retries, catchup and the next backfill.

Learning objectives

  • Run the full four-step review pass end to end
  • Confirm every checklist item on your own pipeline
  • Leave behind a runbook so the next backfill is routine

What you need

  • The DAG you built across this series
  • The pipeline's schedule, SLA and lineage
  • [object Object]

Steps

  1. Apply the idempotency test to every task. (from agent-pipeline-reviewer )
  2. Every task writes to a partition keyed by the logical date, or uses MERGE. (from skill-airflow-dag-reviewer )
  3. Set `catchup` as a deliberate choice, not a default. (from skill-airflow-dag-reviewer )
  4. Retries have a backoff; a task that retries instantly will hammer a downstream. (from skill-airflow-dag-reviewer )
  5. Cap max_active_runs and pool limits so a backfill cannot starve the daily run. (from skill-airflow-dag-reviewer )
  6. Sensors use `reschedule`, not `poke`, so they don't hold a worker slot. (from skill-airflow-dag-reviewer )
  7. Keep pipelines idempotent so a backfill is a routine operation. (from runbook-backfill-required )

Check yourself

  • The four-step review pass — idempotency, retry/timeout/concurrency, backfill prediction, missing tests — runs clean over the capstone DAG.
  • A documented backfill runbook exists for the pipeline, with its idle window and resource budget.

Worked example

The full DAG review checklist — from skill-airflow-dag-reviewer 

1. Every task writes to a partition keyed by the logical date, or uses MERGE.
2. `catchup` is a deliberate choice, not a default.
3. Retries have a backoff; a task that retries instantly will hammer a downstream.
4. `max_active_runs` and pool limits prevent a backfill from starving the daily run.
5. Sensors use `reschedule`, not `poke`, so they don't hold a worker slot.

The four-step review pass — from agent-pipeline-reviewer 

1. Apply the idempotency test to every task.
2. Check retry, timeout and concurrency settings.
3. Predict backfill behaviour.
4. Propose missing integration tests.

Sourced from

Security scan

Last sweep 2026-10-24: 0 finding(s), 0 marked as examples.

Common gotchas

  • A task that appends without a partition key will duplicate rows on retry.
  • Never overlap the production window; never backfill without a rollback path.
  • Backfilling without a written definition of correctness just relocates the bug.

Production considerations

  • Document a backfill runbook per pipeline, with its idle window and resource budget.
  • Keep the fixture of 8 DAGs with seeded non-idempotency bugs in CI so regressions resurface.

Diagram

capstone checklist
   |
   |- idempotency test --> partition or MERGE writes
   |
   |- deliberate catchup --> backoff retries --> reschedule sensors
   |
   |- bounded concurrency --> backfill runbook --> ship

Registry stats

Verification community · Oct 23, 2026
Completeness 81
Trust multiplier ×1.25
Composite 34.15 · adequate
Freshness 1 days since last source check
Difficulty intermediate
Version 0.1.0
License CC-BY-4.0
Security flags none