Upgrade production database to PostgreSQL 12 (latest)

This upgrade will get our application database on the latest production code from PostgreSQL and AWS. It will also serve as the version on which we can begin to build our partitioned tables.

According to SRE, this can be achieved via app-interface.

This will be a database downtime event.

Part of the upgrade procedure will be to take a final snapshot of our database on version 10.
Then the upgrade procedure is executed.
If the upgrade fails, then the snapshot should be restored and reactivated.

Time should be allowed for a full database vacuum/analyze and for a full reindex to ensure optimal performance once the application starts sending query traffic.

Without doing a test run of this vacuum/analyze and reindex, it is difficult to estimate the downtime window. But I would initially estimate 4 hours.

RHCLOUD ticket: https://projects.engineering.redhat.com/browse/RHCLOUD-8422

Deliverables

• Changes to app-interface for the production upgrade
• Addition of script file to vacuum, analyze the database as well as reindex the database.

Notes

• We want to either use the vacuumdb script or have scripts ready to go before beginning the upgrade.