Heartbeat Variance

I’ve noticed in our environment that some groups of computers seem to have a lot of variance even between clients with identical Power Plans and hardware configurations. Below is an example of a computer lab where all clients are the same make/model (Dell AIOs) and have the same power settings, yet online some remain online all the time while most don’t.

I’ve verified they all have their BIOS setup the same as well, and the greyed out clients are definitely online as I can remote into them right away. My first theory was perhaps that Energy Efficient Ethernet was disabling the port when they entered sleep, but disabling that also hasn’t seemed to do anything.

Any advice? It’d be nice to have them all reporting so they run fixlets at roughly the same time.

That gray-out is based entirely upon the last time the client sent a Report. If systems are going into sleep mode/power saving, getting turned off, disconnected from network, or just spending too long evaluating content between reports, the system can become grayed-out.

Whether that has an effect on evaluating new fixlets or processing actions can vary. If it’s just some background evaluation that takes too long between reports, the client can ignore that and give priority to “new” fixlets / actions, executing those & sending a report, before going back to evaluate the long-running things.

Have a look at Tip: Troubleshooting Client Reponsiveness for tips on how to troubleshoot and determine which things the client might be spending too long evaluating. Common issues would be any Analysis or Retrieved Property that descends many folders using the descendant inspectors, or calculates many hashes or large file hashes.

Yep I was aware of the first part. The confusing thing is that all these clients should have the same power settings and I’d expect it to be either all offline or all online but right now I’m seeing a strange mix and I’d prefer one or the other. I know that with Wake on Lan that 3 computers in a subnet will remain online to forward packets, but as my screenshot shows there’s more than just 3.

I can also confirm with your second note that the offline clients do not continue processing actions when they are greyed out (even though they are definitely powered on and are pingable).

I’ll take a look at the responsiveness link you posted to see if there’s anything there I haven’t looked at yet. We have a lot of old Analyses running that my predecessors enabled that I’m sure I could clean up.

Thanks!

“Last Report” does not mean the machine is online or offline - it just indicate when the last Report from that machine has received to the BigFix Server.

I had machines that were online and the last Report was not updating - like @JasonWalker said there are few stuff that can cause this behavior