UPDATE: I have restored 6 days old backup (2025-05-24 at 01:12 UTC). This is not ideal, because there is now 6 days of missing data, and it is my fault I have set weekly instead of daily backups. I would appreciate if you would by any chance had a internal backup that is more recent that you could restore my db into. Please don't proceed with fixes and restorations that would cause service interruptions after 8:30 AM PST, since we have important traffic on our apps after that time. So if it can be done by that time, that would great, if not, please just leave as it is for now.

Thanks!

Unfortunately we don’t have user accessible backups to restore.

Thank you for your reply.

Is it possible to get any logical or physical backup, export, or data directory from the corrupted database state immediately before it was restored with the 6-day-old backup, even if it is corrupted or incomplete.

If it’s possible for your engineering team to provide any of the following from the previous (crashed) instance:

• The raw data directory (even if it’s corrupted)

• A logical dump or partial export from the failed database

• Any WAL (write-ahead log) files or other recovery artifacts

I understand this data may be incomplete, inconsistent, or not usable for direct restore. My goal is to attempt my own recovery or forensic analysis, and I accept all risks.

Additionally, I would appreciate any information or analysis you can provide regarding the cause of the corruption and crash.

Thank you for your help.

We can't do the following on the above, you do have access to the direct container by the ability of SSHing into the DB yourself. However, your restore means that you restored the snapshot overwriting all historical data.

Although I empathize with you, there is nothing more we can do here. I apologize.

Best,
Angelo