Steps to reproduce:

• Use latest versions of server/agents or proxy/agents
• Change the RefreshActiveChecks to 900 then restart the agent
• Using iptables block the connectivity to the server/proxy
• After the agent times out and starts to produce the debug message of “paused history upload” you can remove the iptables rule.
• The agent will now not work for up to 900 seconds

So probably during random data loss for example on bad networks like GSM/Cellular it can happen that RefreshActiveChecks checks every 120/240/... any seconds, connection is alive for 20 seconds, after we get paused state, connection become ok for next 60 seconds, but since RefreshActiveChecks check happens for example every 180 seconds - we fail again, since connection is not alive again.

This can create situation like - don't have any data, due to continuous 'paused history upload' due to the random network instability.