At my previous company when we made the decision to go with Vitess it was because we had confidence that we could accomplish our scalability goals using Vitess for our critical system in a predictable amount of time with a predictable amount of effort and with a safe, incremental, non-disruptive, revertable, testable rollout without downtime.
We had no way to have that kind of confidence about migrating off MySQL to an entirely different database (like TiDB). If we'd been addressing a new storage use case from scratch instead of migrating an existing one, the case for using something other than MySQL might have been stronger.
So it's a story of people who used Kubernetes and MySQL and continued to use Kubernetes and MySQL. The end.