Fast database recovery after application upgrade failure

Question:
What are the recommended (latest) techniques for fast, e.g. < 60 minutes say, database recovery/rollback.

Scenario:
We have a number of databases (application and data schemas) which are impacted when we do software/solution upgrades. Mostly content changes (configuration, data), but some times DDL also.

We take a full DB snapshot/backup prior to upgrade. 

But in the event of an application upgrade failure, data manipulation scripts taking too long, or post the upgrade we discover functional issues, we need to restore all or some of the databases back to the state prior to application upgrade. Data loss at this point is an accepted compromise.

The classic restore from backup, 10s of GB to 1s of TB, takes too long.

Are there MS SQL features, Always on availability groups, or 3rd party tools/techniques that would allow recover in faster time frames, for example < 60 minutes. 

Ideal is zero down time application upgrades but failing that we need to minimise customer down time/impacts. Recovery times become an issue as they're disproportionate to the actual application upgrade time.

Read replications being prompted, transaction database check points/PIT? What does the industry typically adopt here?

Or is the reality that "recovery" is always a timely event (proportional to data volumes and infrastructure).