Update the failback procedure so it applies to production as well as staging
Initially, failback.md
was written strictly with staging in mind. Its quality is lower than failover.md
and it assumes staging in many areas.
I'm dealing with a few cases in !127 (merged) but we should cast a critical eye over the entire document. If we do have to failback in production, our future selves will thank us for putting this work in.
Of particular note are the database failback steps. We should ensure we have high confidence in them.