Your users might report that the system is much slower than they expect. System slowdowns can result from one or multiple problems, including system resources and memory. The tools found in Monitoring & Diagnostics can help you track down these types of problems.

By analyzing the problems described in this topic, you can collect information about which analyses are consuming system resources and memory.

To find the performance counters and analysis statistical information, follow the instructions in Accessing performance data.

• If a Web Player instance indicates high consumption of resources, you can review the Analyses information to determine if you have any particular Spotfire analyses causing these issues.
• If an Automation Services instance indicates a high consumption of resources, then review the running analysis for usage information.

You can get additional statistics for a single analysis in the installed client. You can discover which of its pages or visualizations use most of the resources. See Examining the statistics of an individual analysis for more information.

For information about troubleshooting the data function services, see the following resources.

• Spotfire® Enterprise Runtime for R - Server Edition
• Spotfire® Service for Python.
• Spotfire® Service for R

1. In the list of Performance counters, find the entry for Web Player average processor % for a Web Player instance, or Average processor % for an Automation Services instance. Monitor it for a few minutes.
   • If the entry for average processor % is consistently high, then you have problems with CPU consumption. For a Web Player instance, in the Information area, review the Average load time and Execution time columns. The analyses with the highest values are consuming the most CPU.
   • If the entry for Web Player average processor % is not consistently high and varies, and your performance problems continue, then check for problems with RAM and .NET memory.
2. To check for problems with RAM or .NET memory, in the list of performance counters, review the following values for the following conditions.
   • The value for Total working memory (or Web Player total working memory for a Web Player instance) is high, and the value for Available memory is low.
   • The value for Data engine memory is a large portion of the value for Total working memory (or Web Player total working memory).
   If these conditions exist, then memory consumption is the issue. For a Web Player instance, in the Information area, examine the list of analysis instances. The values listed for the columns Total data table size and Total data view size indicate which analyses are holding the most data table and view memory.

   • If an analysis has a large value for Total data table size, then the amount of raw data can cause problems. Check the analysis to see if it includes any data tables or columns that are not used. If all data tables and columns are used, then you must install more RAM in the Spotfire Server computer to handle the problems.
   • If an analysis has a high value for Total data view size, or if it appears that the number of document nodes is high, the analysis might be too complicated.
     Note: Unused data tables, columns, pages, and visualizations generate more document nodes and use data engine memory. However, unused data engine memory can be paged out to disk when available memory becomes low.
3. To check for additional problems with .NET memory, in the list of performance counters, review the entry for Memory in all .NET heaps. Click Run a full GC(2) twice. This action gives the system a chance to reclaim memory that is released.

   For a Web Player instance, in the Information area, review the Total document node count column. Document nodes are more complicated because they can be different sizes. Analyses that use many more document nodes than the others can cause problems with .NET memory.

   Note: Try to perform this action when the server is not very busy, because the system can be unresponsive while running the GC action.