The data source limits page allows admins to control core query settings for the various data stacks that have been configured on the system. Most of the settings can be set for all queries to all data source types or per data stack technology.
Settings
New Connection Timeout
This global setting determines how long Pyramid will wait to create a connection to any data source.
Pool Size
This setting determines the number of live connections that can be opened concurrently against the data source. This can be set globally or per data stack.
Guidance for Pool Size Limits
Increase the pool limit to execute more queries concurrently against the data source, reducing wait times and improving responsiveness for heavily used dashboards. If the backend has spare capacity, increasing the limit may improve performance further. However, increasing the pool limit also increases CPU, memory, network utilization, and query engine concurrency on both Pyramid and the backend. As a guideline, backend CPU utilization can be driven toward ~70% during peak periods, provided query response times remain stable.
Warning: Beyond a certain point, additional concurrency can saturate the backend and degrade performance. For this reason, you should increase the pool size gradually while monitoring CPU utilization and query response times over peak traffic time. If CPU utilization becomes excessive or performance deteriorates, reduce the pool size.
Query Timeout
This setting determines the time Pyramid will wait for a query response before canceling the query and returning an error. This can be set globally or per data stack.
Fetch Size
This setting determines the size of the data batches that the data stack will send. This can be set globally or per data stack.
Unused Connections Time Limit
This setting determines how long, in seconds, unused connections typically remain alive for the data source before being recycled. This is a "keep alive."