Kai Marshland: for now, you could use cloud storage to upload the data from the worker and download it from the web service. Alternatively, your web service could start a background process to download and process the data when it receives a notification on specific endpoint. Would that work?

Anurag Goel: Yeah, not having shared disks isn't preventing anything from working right now; I've got it set up with a background process.

I'd definitely categorize this as "makes my life easier" rather than "without this my code doesn't work". The reason why it'd be great to have this feature long-term is because the data processing is pretty compute intensive. I wouldn't want to starve my server of cycles, and I could also imagine wanting to scale the data processor horizontally. Uploading and re-downloading is also suboptimal: uploading and downloading adds complexity to my code, it adds dependence on infrastructure other than render, and finally the amount of data (tens of gigabytes each day) makes the time required to upload and download non-negligible.

Kai Marshland: understood. this seems like a special use case (intensive data processing) that requires shared disks; we'll continue to think about how to make this doable without performance penalties (we have some ideas, but we want to learn more about the use cases first).