The server now self-diagnose every 30 minutes for every single unit and reset itself if any problems found. Testing on one of the servers holding 1078 units.
It successfully revived many processes that were killed. And none of them died for a week.
Therefore if DDNS address is pingable, then problem source suspect will be the DVR or Port Digger at the site.
You can find the problem source by restarting the device one by one.
If Port Digger (USB version) has solid LED lights on, I recommend you to restart the DVR first, and the most problems will be solved.
This is why Port Digger needs a separate power source to find out the problem source. If Port Digger USB connected to the DVR, you would not find the real problem source because both will turn off same time. And Port Digger gets the blame.
To prove how sturdy Port Digger is, If Port Digger was using the separate USB power, Try to restart the Port Digger first and let the DVR stays onward. The problem most likely won't go away.
Based on hundreds of support calls, The problem source was the DVR, when the manufacturer is over committing the DVR resources when advertising to compete others (or to make more profit) even when you received knock-off versions,
DVR uptime inevitably reduced and some services like network port, USB port, one of any ports are in the sleep state and would not wake up.
This probably why even your correctly configured manual Port Forwarding requires you to trip many times again.
Once you find the problem source, the next step should prevent it from happening again.
The good news is, our next product will be able to restart the DVR remotely. Stay tuned. We are coming up with the permanent solution this time in very soon.