New 7.5 setup but Inventory keeps failing

We have a fresh install of CMS 7.5 on a Win 2008 R2 server.  I have successfully imported our workstations from AD and pushed the agent to a Win7 SP1 (64-bit) test machine.  But I have not been able to successfully run an initial 'Full Inventory'.  It's been very confusing to troubleshoot, too.  I couldn't find any reports under Jobs/Tasks on the server nor did I see any new Tasks in the workstation agent.  Finally I realized this inventory 'task' was listed under Policies on the server, and that I needed to change the View setting to 'Task Runs'.  This shows multiple failures.  If I view the details for any of them it shows:

Status: Failed
Return code: -2147221005

An unknown exception was thrown. 

CAtrsException exception, error = "Failed to execute task", OS error = 2147746291, at line 388

I googled this error and found this discussion forum!  One of the posts said to look in the C:\ProgramData\Symantec\Symantec Agent\Logs folder and I found a lot of good info there.  The log files show repeated loss of connection to the server.  As a test I disabled the firewall on the server; it's a dedicated box and not Internet accessible.  Also, I keep testing the policy by updating it to run again 20 minutes later.

Unfortunately the connection problems continue:

<event date='01/10/2014 12:51:05.5070000 -08:00' severity='4' hostName='test-workstation-here' source='ScheduleEngine' module='AgentScheduler.dll' process='AeXNSAgent.exe' pid='5360' thread='1420' tickCount='362945935'>

  <![CDATA[A monitored event has occurred, check all schedules]]>

 

<event date='01/10/2014 12:51:13.6910000 -08:00' severity='4' hostName='test-workstation-here' source='NetworkMonitor' module='AeXNetMon.dll' process='AeXNSAgent.exe' pid='5360' thread='3808' tickCount='362954110'>

  <![CDATA[IPv4 address change detected]]>

 

<event date='01/10/2014 12:51:13.7330000 -08:00' severity='4' hostName='test-workstation-here' source='NetworkMonitor' module='AeXNetMon.dll' process='AeXNSAgent.exe' pid='5360' thread='3808' tickCount='362954156'>

  <![CDATA[Server down [0x00000000]: https://internal-server-addr-here:443:{B81D1ABF-D418-4561-B9A5-BB71CEA8AE84}]]>

My suspicion is that our self-signed SSL cert fails intermittently but I don't know how to prove that.  It works at least some of the time as the workstation is reporting 'basic info' as well as receiving task updates; also I can manually open the HTTPS URL in a browser.

Any suggestions?  I greatly appreciate it!