BUG: SQLAlchemy error

[zzzepezzz]

I got many 500 internal server error when navigating throw the Aleph UI.

I found that sometimes, when the postgres stack reboot, it may cause the SQLAlchemy session to be invalid.

Precision, I only reproduce this bug using an aleph stack with oauth

Aleph version
Aleph v4.0.2

Discourse subject

The only way I found to fix this error is to add the option pool_pre_ping in the SQLAlchemy initialization.

[zzzepezzz]

With some debugging, we just found that this solution fix our problem (issue #4002)
Solution: #4002 (comment)

Just set your ALEPH_DATABASE_URI and FTM_DATABASE_URI env vars (even if they just point to the defaults)

[zzzepezzz]

It seems after multiple days of tests that #4002 is not the solution. I keep having several 500 errors