1. Get physical access to the systems; employee badges were also offline, so they needed to bring in a different set of people with physical access to the premi

1. Get physical access to the systems; employee badges were also offline, so they needed to bring in a different set of people with physical access to the premises. 2. Login directly to their systems and update the A-name record. 3. A-name records do not take effect immediately. There's a propagation delay as other DNS services accept the domain name. 4. Any sort of "damage assessment" to check the conditions of the now-offline systems. [Not a gating factor; you can turn things on, but you are likely doing "blast radius" analysis in parallel.] 5. Restart services one-by-one. Web servers, databases, API gateways, etc. 6. Allow external access; turn the systems back on for the world. [This is where they are currently, even though some parts of Facebook do not seem to be working too well. i.e., uploads and new posts seem broken still; at least they were for me.]