Tracking down source of periodic CPU spikes in Orthanc 1.5.3

I recently updated a number of my Orthancs to 1.5.3 from older versions (ex. 1.2.0).

I noticed on occasion that there are CPU spikes of activity, ranging 20-50% CPU usage every 8-10 seconds on some, but not all of my updated Orthancs. Some that exhibit the CPU usage spikes have no DICOM.

I’ve been trying to track down the CPU spikes, thinking maybe I’ve configured things incorrectly for the new version.

Might these be part of the new jobs engine? Is there a polling period setting set somewhere?

These Orthancs all have postGRE backend databases with DICOM stored on disk (no storage in the db). Would the DB communication be causing these periodic spikes?

I thought it might be the background check on stable studies, but as I mentioned, one of the Orthancs with no DICOM also exhibits this periodic CPU usage spike every 8-10 seconds.

I tried running with debug mode turned on, but did not see any particular reports in the log that would correlate with the timing of the spikes.

Any suggestions?


Please upgrade to Orthanc 1.5.6 and PostgreSQL plugins 3.2. These releases fix performance issues.

Also, you could try and set the “SaveJobs” option to “false”. By default, jobs are backed up every 10 seconds in the database.

Question: what version of orthanc is in the jodogne/orthanc-plugins docker image?


You can use the “jodogne/orthanc-plugins:1.5.6” image, that contains the latest official releases for the time being (Orthanc 1.5.6 and PostgreSQL plugins 3.2). For instance:

$ sudo docker run -p 4242:4242 -p 8042:8042 --rm jodogne/orthanc-plugins:1.5.6

More information is available in the Orthanc Book:

Click on the “Plugins” tab of Orthanc Explorer in order to see the version of the plugins.

Thanks for the quick response, Sebastien. I’ll probably try this setting since I haven’t yet made use of the jobs facility directly.