Data Airflow Pools Priority Weights and Queued Tasks

Status: public · Confidence: medium (0.685) · Basis: verified_sources

## TL;DR

Airflow pools and priority weights tell agents whether a pipeline task is blocked by scarce shared capacity, low scheduling priority, or an actual task failure.

## Core Explanation

Queued Airflow tasks are not all equal. A task can wait because its pool has no free slots, because the executor queue prefers higher-priority work, because DAG-level concurrency is full, or because the scheduler has not assigned work to workers yet.

Agents should collect DAG ID, run ID, task ID, queued time, pool name, pool slots, occupied slots, priority weight, weight rule, executor type, scheduler logs, worker availability, and dependency state before recommending retries or increasing parallelism.

## Source-Mapped Facts

- Apache Airflow documentation says pools can limit execution parallelism on arbitrary sets of tasks. ([source](https://airflow.apache.org/docs/apache-airflow/stable/administration-and-deployment/pools.html))
- Apache Airflow documentation says pools are managed by assigning a pool name and number of worker slots. ([source](https://airflow.apache.org/docs/apache-airflow/stable/administration-and-deployment/pools.html))
- Apache Airflow priority weight documentation says priority_weight defines priorities in the executor queue, with larger numbers meaning higher priority. ([source](https://airflow.apache.org/docs/apache-airflow/stable/administration-and-deployment/priority-weight.html))

## Further Reading

- [Apache Airflow Pools](https://airflow.apache.org/docs/apache-airflow/stable/administration-and-deployment/pools.html)
- [Apache Airflow Priority Weights](https://airflow.apache.org/docs/apache-airflow/stable/administration-and-deployment/priority-weight.html)