Hacker News new | ask | show | jobs
by ajbosco 2616 days ago
I thought they used Piper (based on Airflow) for workflows at Uber. https://eng.uber.com/managing-data-workflows-at-scale/
1 comments

That is for data workflows (your etl jobs basically). This is for services.
If you already have a service that requires ETL, Cadence might be a good choice as well. You will need more boilerplate and setup compared to Airflow/Piper, but you can share data structures between ETLs and services that use it and have more control over execution and deployment as you own your workers.