Native database backups for popular databases

rc

a month ago

It'd be great to not have to worry about backups on Railway at all.

We offer an S3-backed automated Postgres backup template, but having this in dashboard itself as a tab would be killer.

  1. Input S3/GCS/some other cloud storage credentials
  2. Set a schedule
  3. We automatically
    pgdump
    /
    mysqldump
    /etc. your database to the configured cloud storage at the schedule you configure

"Popular" meaning widely-deployed on Railway, e.g. Postgres, MySQL, and Mongo. A more generic way of doing this would be automated volume snapshots that we can ship to any cloud storage provider, and/or combining the two (volumes -> ship snapshot; databases -> ship native dumps).

5 Comments

👆 would be amazing!

I think going for both volume, and "popular" databases would be the way to go, you don't want to lose

pgdump
as an example. We also run other databases such as Chroma vector store, which would be cool to backup as well.


faces-of-eth

a month ago

I'd add Redis to the list perhaps but echo the original poster's request.

Additionally, being able to load this data, or a seed data file into PR environments or forked environments would be clutch. Right now I'm hacking this as a custom one-off data loader service that only runs in environments lacking a PRODUCTION environment variable


dwaynemac

a month ago

maybe this could be implemented in a more generic way providing a native backup solution for VOLUMES.


brody

a month ago

the team has mentioned volume snapshots as the backup method, this would allow taking backups of services that use sqlite databases, along with so much more


dwaynemac

a month ago

Volume snapshots would be awesome