UCA Agent failing to connect with "Not Connected" or "Never Connected"

[Krishan]

Hi 

I am trying to set up the IR UC assessor on our network, we have opened up the neccessary firewall ports and defined the proxy server addressing as recommended in the start up guide, however we are still failing to get the client failing to connect.

We have an infrastructure that uses Zscaler as our proxy and using Windows 10 Dell laptops.

Is there any troubleshooting steps I can go through to investigate further.

Warm Regards

[GeraldC1 · Accepted solution]

Hi Krishan,

I received good news that you resolved it with proxy configuration for the required secure websocket connection. Here are some general troubleshooting steps for this kind of scenario: 

 

1. Confirm which agents (or all of them) and is it "Not Connected" or "Never Connected";

2. If "Never Connected" then make sure the agent is installed (on the designated agent PC Download Installer, copy to clipboard the Key, run the installer );

3. If "Not Connected" then check if the 'Prognosis Agent' service is running on that PC, if not, start it up.

4. Consider Help Center: 'Firewall requirements' and 'Configuring the Prognosis Agent to use a forward proxy' articles;

5. Test successful initial connection in a browser (not IE) this 'https://data.iot.us-west-2.amazonaws.com/' returns: 

{"message":"Not Found","traceId":"93f9…"}

6. Test successfully kept connection over several minutes, in a browser 'https://www.websocket.org/echo.html' with wss://echo.websocket.org 

7. If these work from browser, check the browser proxy server setting, set the same for the agent in the PROGNOSIS_HTTP_PROXY environment variable using Help Center 'Configuring the Prognosis Agent to use a forward proxy'.

8. Check for 'error:' in the log "C:\ProgramData\Prognosis Agent\Logs\agent.log" if still failing and Google search the error messages, as they can help identify issues in the environment outside of the Prognosis Agent.

9. Submit a request for assistance to IR and attach log and details.

HTH

[GeraldC1]

Also try give the Prognsis Agent service a restart and then check if the service started up and spawned processes successfully and then was able to establish a TCP on port 443 connection to the amazon cloud:

tasklist | findstr "PrognosisAgentService.exe node.exe Session# ========"
netstat -ano | findstr "<the PIDs identified in the tasklist output>"

Example:

When the Prognosis Agent has connected successfully:

Another example with comments explaining: