High CPU Usage Causes DICOM Web API Slowdowns and Server Crashes on Orthanc PACS

We are currently hosting our Orthanc PACS server on an AWS EC2 instance with the following specifications:

  • vCPU: 8
  • Memory: 32.0 GB
  • Processor: Intel Xeon Family Processor at 3.3 GHz

Our setup also includes the mysql and s3 plugins for storing DICOM studies. The primary use case for our PACS server is to perform microscopic zoom operations using DICOM Web APIs. The DICOM study file sizes we handle typically range from 3 to 6 GB.

Issue

When zooming and scrolling around the image, the CPU usage spikes up to 200%. During these spikes, the web API becomes unresponsive and extremely slow. This high CPU usage often causes the server to become unstable, necessitating a full restart.

Request for Help

We are seeking advice on how to optimize our PACS server to handle these operations more efficiently. Specifically, we want to:

  • Reduce CPU usage: Ensure that CPU usage remains within acceptable limits during intensive zoom and scroll operations.
  • Improve responsiveness: Maintain API responsiveness even under heavy load.
  • Enhance stability: Prevent server crashes and minimize the need for restarts.

Any recommendations or best practices for optimizing Orthanc PACS server performance in our use case would be greatly appreciated. Thank you!

Orthanc DICOM Server/Client v.1.12.3 with plugins

Deer Orthanc Team

Please follow up on my previous inquiry regarding an urgent performance issue we are experiencing with our Orthanc PACS server.

Given the critical nature of this issue, we would greatly appreciate any recommendations or best practices you can provide to help us resolve this problem. Your timely assistance and expertise are highly valued.

Thank you for your attention to this matter. I look forward to your prompt response.

Hello,

If your issue is critical, you should consider purchasing a support pack.

Hi,

As Benjamin explained, you should not expect to get immediate dedicated/critical support on this discussion group - at least not from me :wink: .

The first thing to do when you have any kind of issues with Orthanc is to analyze logs in verbose mode to understand what are the API routes that are being called, and where is Orthanc spending time or using CPU.

In general, also check how to request support.

Best regards,

Alain