An issue with database replication for our DNS nodes was causing some stale records to be provided. Any DNS record lookups on SystemDNS name servers would intermittently respond with stale results, depending on which name server was hit. The results would be the records from before the issue start time Thursday (July 30th).
Our engineering team was able to resolve the issue with a copy of the database.
Incident Start Time:07-30-2020 19:20:00 UTC Incident End Time:08-01-2020 08:15:00 UTC Total Duration:35 hours 55 minutes