For compliance we need to maintain native SQL backups. On-prem systems use Veeam backup which easily gathers daily SQL backups and manages a retention policy of daily for 2 weeks, end of month for 6 months, end of year for 7 years. The company is now building the next technology stack on Amazon Web Services and the S3 bucket version management is woefully simplistic. I have created a lambda to trigger a native SQL backup once a day, and hoping to be able to manage the version retention, it simply overwrites the file in S3. This DB is expected to grow to around 4TB by the end of the year, so paying to store every version every day for 7 years is out of the question.
Has anyone in this group come across, or written a lambda (or other widget) that can be triggered to look through S3 previous versions and prune the excess according to a selected or defined retention policy as described?
You can lease around 12TB for <$90 USD/month. These are Linux machines, so no upload/download bandwidth charges. No per object access charges. No other charges. Just one flat rate.
You can simply create versioned tarball files each midnight, pruning by number of days or maintain some percentage of free disk space.
Tip: You can also merge the two approaches, pulling a copy of your S3 bucket(s) each midnight to your OVH machine, then generating versioned tarball files.