A case study in how not to deploy software in 2023: What was supposed to be a 12-hour downtime for a launch of a total rewrite of their mobile app and backend services turned into a 48-hour full outage. Six days on, many users are unable to sign in and view or modify their accounts that are still being shown to other users. The company is not responding to support requests except with a canned response with no timeline for affected users.
The outage deserves attention due to the severity, length, and privacy implications for users of dating apps. I would love to see the postmortem starting with how they decided to do a total rewrite of everything and launch it as a big bang release of a new backend and client during a planned downtime window.
Reddit reports of ongoing outage: https://old.reddit.com/r/feeld/comments/188m4c9/update_discussion/
Twitter announcement of downtime window Nov 30: https://nitter.net/feeldCo/status/1730144506555642103#m
Twitter acknowledgement of still unresolved partial outage Dec 3: https://nitter.net/feeldCo/status/1731141992132800652#m