Add pool_pre_ping to engine creation options

[meksor]

Another stab at the postgres connection timeout problem! This is kind of a hack but pool_pre_ping sends a "ping query" up to three times before trying the actual query, should help with our idle-connection problem at a small cost until this is figured out on the setup side.
https://docs.sqlalchemy.org/en/20/core/engines.html#sqlalchemy.create_engine.params.pool_pre_ping

[glatterf42]

Looks good to me :)

What does it mean, though: "until this is figured out on the setup side"? Should ICT do something? Do we need a tracking issue for this?