The VZ Socrata ETL DAG updates both the crash and people datasets all in one task. It could be beneficial to break these crash and people dataset replacements/updates into two separate tasks. The crash dataset takes much longer to run. If an issue occurs during the people dataset update, the whole DAG must restart and could delay patching the datasets to get Vision Zero Viewer back online (since it is powered by this dataset).
The VZ Socrata ETL DAG updates both the crash and people datasets all in one task. It could be beneficial to break these crash and people dataset replacements/updates into two separate tasks. The crash dataset takes much longer to run. If an issue occurs during the people dataset update, the whole DAG must restart and could delay patching the datasets to get Vision Zero Viewer back online (since it is powered by this dataset).
See for new info about re-running individual tasks: https://austininnovation.slack.com/archives/CHZE6BC6L/p1690903981451939