Drawbacks
The main drawback is failure at scale. Code updates can be problematic, especially given Drupal’s value proposition is its extensibility. Updates that introduce regressions are not caught on just one or two sites, they're deployed to all of them. And, once the code is updated, the Drupal application needs to be updated. This must happen immediately. It is impossible to manage failures in just one site; each site needs to be remediated. And, it’s common that a code-level failure applies to one if not all, sites.
To do any level of testing before deploying to a large number of sites, it is important to use environments and perform rigorous testing against multiple sites. Check logs for anomalies and run automated tests. All of these steps should be done before a production deployment to mitigate risk.
A similar approach is leveraging a shared codebase with orchestration but without multisite. Developers often use Git repositories to manage Drupal codebases...