I have four sites with current firmware BRXs sending daily production emails to management. Two of the sites were flashed in February to current firmware to take advantage of adjustable TLSTimeslice. One just came online late June. Empirically I found a time slice of 3000 or so worked reliably, so I set at 4000 to give some margin. All four sites ave been working well, using the same GMail account and the same BRX program (except for one site with slightly different PLC).
Two of the sites spontaneously quit emailing late last month. I looked at one, but the only time I had to troubleshoot, all I learned was that the fail relay was coming on, indicating failure to EMAIL. Now, I have local PC's at three of the sites. So for the one that quit working but had a PC I remoted in and ran DMLogger, only to find that GMail was dropping the connection after 10 seconds. Found I could make that one work by increasing the TLS time slice to 10000 (when 4000 worked before). No change in the PLC program or scan time.
So I assumed the other failing one presumably had the same (inexplicable) issue. Logged on remotely today from my own laptop, since there's no local PC, and even a time slice value of 50000, while it does increase total scan time, does not result in a successful email. The scan time also seems to have a sawtooth of maybe 2ms with a period of maybe 4.5 seconds. The screenshot was taken with a time slice of 10000, which it's using, but never successfully emailing. If I test creds from DMD device configuration it succeeds. The creds should not have changed since it was working, and the same login and password are in use at the three working sites (including the one where I increased the time slice).
This is weird. I can't figure out what it might be other than something with GMail, but can't get DMLogger info for this one particular site, natch.