Heavy slowdown et freeze issue


Since a day or two, we notice again heavy slowdowns on our application. Randomly, some models are not displayed at all (white screen) or remain ‘blurred’…forever.
Overall, notable slowdowns affect all our models.

The problem seems to impact more especially models that have been previously saved, but the display of blank models is also very slow.

Here is an example of a “cutomize” request that is followed by a “cache” request loop that never ends:
fetch example.txt (6.1 KB)

Have you done any particular update, which could have consequences on the cache or the response time?

Once again, this problem has heavy consequences on our productivity and the experience of our customers.

Thank you in advance for your help

Hi @julien.chilou, no updates have been rolled out in the past month until today about 2 hours ago, which can’t possibly explain the impact you are experiencing.

Please explain what you are referring to precisely by “models that have been previously saved” and “blank models”. Does this refer to models newly uploaded to ShapeDiver?

We have rolled out an update today morning, which should go unnoticed except for general performance improvements (less cache requests, customization requests typically waiting until solution is finished, more consistent response times). Please let us know whether you notice any difference.


please find attached. a “customize” request for which there is an infinite call of “cache” request. (captured in chrome devtools)
Hope that it could help.fetch example.txt (6.1 KB)

thank you

The models I refer to are the models of my customers: the soles and not the GH definition which have not been modified. Nothing has been modified on the GH side for several weeks.
Concerning the ‘blank models’, I mean the new products, i.e. without any customer parameters and without importing any external 3d file. That is to say what is displayed when one of our customers starts to make a confection.
Our workflow is as follows: the patient chooses a product model (i.e. one of our GH definitions registered on SD). This is loaded and displayed ‘flat’. The customer loads a 3d scan file and the sole is shaped on the foot. Then he has many parameters to modify the sole according to his needs. Finally he saves the model of his sole and proceeds to the order. The registered model can be reopened at any time, the parameters of the customer being preserved.

To return to our problem. I can create a new confections that seem to work well, slowly but nothing out of the ordinary. And for some (not all) of the confections made by our customer for example one or two days ago are impossible to open.

Many thanks for your explanation @julien.chilou. If I understand correctly, the first part of your workflow works fine, but when a configuration that was previously computed (and should be cached) is recalled, that causes an endless loop of cache requests. Is that correct?

Yes. That’s it.