[FEATURE] Separate Public & Private S3 Bucket

[jonas-hoebenreich]

Is your feature request related to a problem? Please describe.

Currently, the same S3 bucket is used for both private data (file exports) and public assets (uploaded images). This approach has several drawbacks:

• Security Risk: Private data should ideally be deleted after the download is completed to prevent unauthorized access. Using the same bucket for both private and public data increases the risk of data leakage.
• Performance and URL Management: Public assets would benefit from being served via CloudFront with a clean URL and efficient caching.
• Adding a Cloudfront distribution could potentially leak files from the /private folder.

Describe the solution you'd like

1. Private Data Handling:
   • Store private data in a temporary folder within the Docker container.
   • Ensure this data is deleted on container restart to reduce the chance of data leakage.
2. Public Files Management:
   • Serve public files via a separate S3 bucket.
   • Allow users to configure this S3 bucket to use CloudFront for distribution to eliminate the need for authentication and to implement an efficient caching policy.

Benefits:

• Enhanced security by segregating private and public data.
• Improved performance and user experience with CloudFront's caching capabilities.
• Cleaner URL management for public assets.
• Browser starts caching assets -> reduced bandwidth fees.

Describe alternatives you've considered

No response

Additional context

No response

[mattinannt]

@jonas-hoebenreich thanks for raising this issue :-) With private files you mean e.g. the temporary download link for the responses download? Or also the files provided by survey participants as file uploads?

[jonas-hoebenreich]

Ah, good point. Do be honest I don't really care about survey participant file uploads, I guess the current approach does work for them. Maybe just rename the files to something random to avoid that links can guessed. My concern are mainly response exports that contain sensitive data.