I finally fixed our staging DNS mess at the Austin shop, took 11 days instead of 2
We had a wildcard CNAME that silently masked records for half our staging services, so nothing resolved when we moved to a new VPC. I figured it was a quick records fix, but every change took 30 minutes to propagate and nobody kept a change log, so we were retesting blind for 11 days. Ended up rebuilding the whole zone in Route 53 from a hand-drawn diagram at 2am, and it worked on the first try. If you run staging on a wildcard, go check it now before a deploy eats your week. Anyone here keep a real change log for DNS, or do you just rely on memory like we did?