As a data guy, I've have an entirely different perspective on serverless. While it may not work well for running entire parts of an application, it works great

As a data guy, I've have an entirely different perspective on serverless. While it may not work well for running entire parts of an application, it works great for building out and managing internal data pipelines. You don't want your pipelines crashing because your jobs overlapped and you ran out of memory/cpu on the servers you set up. You want each stage of the pipeline to run independently and scale as needed. In my eyes, the bigger problem with many FaaS products (which most equate with serverless) is the barrier to entry. The setup isn't very user friendly. The code doesn't "just work" like it does locally. Limitations on data size, runtime, etc. cause you to building workarounds in your scripts just to get them running. Not to mention that once you have it all set up, visibility into everything running is a nightmare and only available to the most technical users. Based on my experience, I'm currently building a [platform](https://www.shipyardapp.com https://www.shipyardapp.com) to try and make serverless data pipelines easier for teams to setup and manage. Would love to hear someone else's perspectives on serverless setups for data management. I know I'm not alone on these existing frustrations.