GitHub user matrach added a comment to the discussion: Scheduler performance with large number of mapped task instances
I've mentioned the mini-scheduler above, but if I understand correctly, it resolved the scheduling state for ALL task instances, but then ignored the ones not downstream...? It appears that the scheduler for mapped tasks takes at least quadratic time in their quantity, as [here](https://github.com/apache/airflow/blob/577985821e4e3c8cfbc2797feb19cc32fe2aaac0/airflow/ti_deps/deps/trigger_rule_dep.py#L255) it loops over all finished task instances, and does so for all schedulable task instances. GitHub link: https://github.com/apache/airflow/discussions/46044#discussioncomment-11954511 ---- This is an automatically sent email for commits@airflow.apache.org. To unsubscribe, please send an email to: commits-unsubscr...@airflow.apache.org
