Investigate putting volume snapshots in S3
Some services we deploy require a persistent volume for local files, which needs to be backed up. There's two way to make such a backup:
- Have a cronjob that puts an archive in S3. Good for daily backups of small volumes, cost efficient, longer to set up.
- Have a cronjob that calls the scaleway API to trigger a volume snapshot. Good for less frequent backups of large volumes, more costly, quicker to set up.
It is also possible to push the created snapshot to an object bucket afterwards. We should investigate to see if it provides advantages in terms of:
- cost efficiency
- ease of setup
- ease of restoration
Other ideas of how to improve backups are of course welcome.