Cloud server times fluctuating wildly?

I found a couple old threads about this error. I started getting it yesterday from DMT, 5 times in 50,000 records.

2022-03-18T02:24:35 PartNum: EST0053644 The security timestamp is stale because its expiration time (‘2022-03-18T03:54:48.839Z’) is in the past. Current time is ‘2022-03-18T09:24:35.597Z’ and allowed clock skew is ‘00:05:00’.

Note that there are several hours between the expiration time and current time. But it’s not our UTC offset and the minutes are also off by varying amounts, so it’s not a time zone thing. (The difference between the log timestamp and the “current time” in the message is our UTC offset.) So either the server didn’t notice for several hours that the client auth was expired, or the server time keeps changing by several hours. The client time is correct and steady. We’re using time.windows.com for NTP. Our Appserver URL is blank.

Any other cloud users seeing this?

I was updating Part with Company, PartNum, and InActive. The parts were test parts that had never been used for anything. All rows that got errors went in without errors the second time. Nevertheless, support automatically blamed my data.

I also got errors about rows being modified by another user over the weekend when no other user was on. After implying again that it was a problem with my data, support said this could have been caused by a backup process running.