Tune the Model Repository Service to improve performance. You can configure the Model Repository Service process property for memory and minimize network latency.
Consider the following solutions for Model Repository Service bottlenecks:
Host the Model repository database on the same machine as the Model Repository Service.
The Model Repository service process performance can be affected if the Model repository database is hosted on a remote server. Model Repository Service operations that require communication between the Model repository and the Model Repository Service on a high-latency network could slow down the Model Repository Service performance.
Increase the Maximum Heap Size property for the Model Repository Service process.
The Model Repository Service process can consume large amounts of memory while processing a large number of concurrently logged in users. This can cause a large number of network connections to be open between the Model Repository Service and other services such as the Data Integration Service or the Analyst Service.
Use the Administrator tool to configure the Maximum Heap Size property to a larger value in the Advanced Properties for the Model Repository Service process.