Heartbeat Stops Reporting Knocking PC Off Network

Oh I should also add before someone asks, when Sophos Central reports the device as no longer sending a heartbeat, I do acknowledge the event to allow the PC status to go back to green because my firewall rule does have the Security heart enabled on the PC must at least have an amber status or greater to maintain its network connection.

Heartbeat can periodically not be sent when services are starting up/recycling or any other network interruption. Also, at reboot, the services don't start in a set order so the network elements could start up before Health does and therefore there is no Heartbeat to send. 

In general, the endpoint itself isn't stopping the traffic - it is the firewall that is dropping the traffic. Is this traffic bound for an external resource or an internal one?

Our topography would be considered the standard star topography. The firewall, dhcp server, PC's all communicate amongst themselves.

I have turned off Device Isolation to monitor over the next few days, I assumed it was necessary to have this enabled, in case of a PC possibly being compromised and then preventing any lateral movement through the network.

try turning that off and see if the issue goes away - if it does then you might want to consider re-eval of that policy. In essence, it is meant for high priority machines/network segments and when you turn it on you are saying you accept the risk of interruption to service because you are more concerned about lateral movement of a compromise to high-priority systems.

Thank you RichardP, really appreciate your input. I will see how it goes til midweek because I  normally would see the condition on at least a couple of PC's between then and now.

yes - that is exactly what it does. The balance comes though on how much your prioritize that protection over network availability. With Isolation turned on - you will see service interruptions from time to time. 

Thank you, I do have my firewall LAN rule requiring an amber heartbeat so if it something is compromised and changes the status threshold, that should still provide me some sort of reliable security laterally but I shall see.

Think of it this way:

SFOS will protect from infections spreading between network segments that have to transverse it to talk to each other.

Isolation protects from infections spreading inside a network segment.

However just a quick note, in theory with device isolation, if the administrator acknowledges the device has gone into this red status, should that now allow the PC to resume communication after its status has been reviewed and deemed no threat? Is that not the purpose of reviewing an issue and then acknowledging it? Because on Sophos Central the device then changes from red status to green status, the software on the device changes from red to green but yet it still never recovers, seems to be a bug no?

Isolation only goes into effect when the Health on the machine is RED. So, once you complete the action that turns it to GREEN, the isolation will be removed on that machine. You can confirm in the registry: Computer\HKEY_LOCAL_MACHINE\SOFTWARE\WOW6432Node\Sophos\Health\Status
The admin, health, service, and threat values are what you care about. 1 is good. 3 is red. So, you can check what the machine thinks it is.

Since you have two layers that can be affecting network traffic you might be encountering a collision. So, endpoint turns RED, it isolates and reports to the SFOS that it has RED health. SFOS also blocks. If the SFOS is safeguarding the comms to the DHCP server too - this could be a problem. 

Do you have local DHCP servers in each segment? Or are they all beyond the SFOS from the perspective of the clients? Like you have the DHCP server in segment A and all machines have to route through the SFOS to get there?

Everything on site is LOCAL, the XG being part of the local LAN network along with the DHCP server. XG is our gateway/firewall device, if it were removed the devices would simply not have a way out to the internet, that is it. Should this occur again with device isolation off I'll be sure to check the registry key status.

Everything on site is LOCAL, the XG being part of the local LAN network along with the DHCP server. XG is our gateway/firewall device, if it were removed the devices would simply not have a way out to the internet, that is it. Should this occur again with device isolation off I'll be sure to check the registry key status.