Hello,
We deliver Reporting Services reports to our customers via a web application with an embedded ReportViewer control. We have 3 Reporting Services servers load-balanced to provider high availability and increased capacity. During peak usage, the SSRS APIs (http://hostname/reportserver) become periodically unresponsive for tens of seconds, causing a poor user experience. The hosts that run the unresponsive services are fairly underutilized during this time (< 30% CPU and no memory pressure). Each server has 16 CPUs and 1.5 TB of RAM. It feels as if the system is queueing requests, but there is nothing in the logs that says it's doing so.
I have tried changing various rsreportserver.config configuration settings in an attempt to increase parallelism, but haven't found anything that resolves the problem. Here are my non-default config setting:
<Add Key="MaxActiveReqForOneUser" Value="100"/>
<Add Key="MaxActiveReqForAnonymous" Value="500"/>
<RecycleTime>10080</RecycleTime>
<WorkingSetMinimum>20000000</WorkingSetMinimum>
Is there a configuration setting that I can change to allow the service to handle more reports or connections concurrently?
Thanks,
Mitch
SQL Server ReportServer database Version: 14.0.3208.1
Reporting Services Version: 14.0.600.1274