Problem:
Spotfire can often indicate a node is strained or exhausted, but also indicate that CPU and Memory are OK. To know why, you will need to look in the log files.
Details:
You can access the Node Manager logs, via 3 different methods.
- Remote into the Node Manager's desktop (Windows) or use SSH (linux), and open the \tibco\nm\nm<version>\logs folder.
- Download the logs via a the Spotfire Admin Monitoring & Diagnostics page. Click the 3 dot menu at the end of the Node Manager line, and select Download troubleshooting bundle
- View the logs via the Spotfire Admin Monitoring & Diagnostics page. Click the View Logs link.
Look for the Spotfire.Dxp.Worker.Host.<GUID>.log file. Open this and find any lines that contain "instance health status". These lines will indicate all CPU, Memory, Paging, TempDisk, WCF, ThreadPool and Combined status.
Paging: How much data is being pushed to and from disk, because of a lack of memory.
WCF: Microsoft Windows Communication Foundation. These are communication channels between software components. This is an indication that the system is not too busy.
ThreadPool: Is another indication of how busy the system is.