I’m experimenting troubles with the server logging:
My code used to log a measure around one time a second. And now it is doing it erraticaly, it could spends 10 or more seconds without logging anything and after, log all the measures, causing a message throttling and the consequent disconnection of the imp.
Logs from imp or logs from agent? Do the timestamps show all the same, or once per second and just the IDE window is updating in blocks? Anything interesting about the internet connection the imp is on?
Oh, and if this is on the imp, any sleeps in your code?
When you say “sleep routine”, what do you mean? It calls imp.sleep? When you do this, the TCP stack doesn’t run (as the squirrel runs at a higher priority), which could cause buffered messages not to escape the imp.