No output log that reveals too much network issues. Devices seem to idle little, at least 2 from 4 units but no log output. I guess that's BFL hardware based scaling down because of HW error count. I must install network monitoring tool like prtg, changing to picuntu linux anyway but seems more this is a perhaps BFL D-grade chips caused fluctuation by temperature perhaps. 3G connection ping & up bandwidth is not too great or stable, that can perhaps affect too. Also don't know should I use higher difficulty setting like -d16. I noticed that bitminter does that dynamically so I didn't use any forced option. Been GPU mining so these ASICS & stratum protocols new to me.
The difficulty will go up automatically.
Do you see on the server that workers are idling, or on the client side? Maybe you see only hardware errors on the client side (and a hashrate), but zero hashrate on the server side, because no results are accepted?
Look for hashrate going to zero on the client side, and look for network errors or "out of work" type of errors in the log.
Webpage has been offline for a while, mining is working normally.
I botched an upgrade and it took 5 minutes instead of 5 seconds. Sorry about that. It also looks like the website can still respond slowly at times. Still working on that.
I also moved over to a forced HTTPS setup, trying to eliminate HTTP as much as possible. Let me know if there are any issues. Downloading the client and accessing the API is still possible over HTTP, otherwise that might break things for people. Maybe eventually it can be 100% HTTPS.