Nextstrain / Github LFS hosting / Bandwidth Issues

1

We’re currently using Github to host our weekly and historic Nextstrain builds. Some of these builds have grown in size substantially, so we use git LFS (and Github’s associated LFS storage) to handle these. Recently we keep running out of bandwidth (as determined by Github’s LFS quotas) despite increasing our quota/paid resources. Base on this, we were hoping to get clarification on the following:

  • If we host multiple builds on our Github repo via LFS, will all of those builds be downloaded each time someone loads our Nextstrain community build, regardless of which specific build they’re looking at? Or does it only download the one currently being viewed?
2

Hi Rebecca – each request for a dataset only fetches the files of the dataset (the main JSON, plus any sidecars such as tip-frequencies), however using GitHub like this is probably pushing the limits (and costing you). I would suggest exploring Nextstrain Groups which scales much better - let us know if you’re interested!

3

Great, thank you for your quick response! I’ll take a look at this option.