People & processes
Good article.
Might have been worth touching on people and processes rather than just tech. 90% of the instances I have seen requiring restore from backup have been due to fat fingers by a sysadmin. Most notably the guy that took a copy of the Prod database to a test system and then anonymised the prod database instead of the copy. OK so backup saved the day but it took two days to restore that backup and get it up and running.
1. Restrict write permissions to Prod by exception (break glass proocedures for admin access to Prod)
2. Have 2 or 3 eyes on all changes involving production data by default (even if they appear low risk)
In addition its worth noting that if you are using public cloud for resilience then you need all of your security apps in there as well (DLP, SIEM, etc). Obviously theres a cost/benefit analysis to be done, but if your private DC has gone up in a puff of smoke then you may be running a huge security/ compliance risk if you are running out of a Public cloud with no security.