Create a blueprint on offloading large objects to secondary storage
While git is designed and optimized to track the changes of text-based files, we start to see the trend that our customer store large files in git for various reasons.
Git does not have a native way to handle large blobs well. Once a large file ( an image or a video) is added, it stays in the repository history. The repository size is quickly bloated.
We want to design an architecture with promiser origin setup to offloading packfile with large object to cheap object storage.
- Large file packfile can be offloaded from server to object storage to save storage on server
- Files should be sent direct from storage to the client to save network traffic
- Minimum client setup to avoid configuration complexity.
Edited by Eric Ju