Unable to send a study with large number of runs through API

[prateekwahi]

Hi!

I have been conducting large batches of simulations using the API, with minor adjustments to the scripts provided in https://discourse.pollination.cloud/t/unable-to-load-a-large-number-of-runs-as-one-job/1446/9?u=prateekwahi.

However, since yesterday, I have encountered issues uploading 1300 samples. I checked, and it worked with 100 samples, but as soon as I attempt to upload all 1300, Python returns a “403 Client Error”. I have verified the project name on the cloud, the Python script, and the recipe “custom-energy-sim 0.3.19”.

It would be greatly appreciated if someone could look into this, as I have conducted many simulations with large batches of samples successfully.

[mostapha]

Hi @prateekwahi,

1. Can you remind me your cloud subscription tier? We started to enforce the number of runs per study based on the subscription after some users submitted studies much larger than the number of available parallel CPUs which resulted on large number of studies being in queue for hours. You can find the number on the Compute Package Subscription card.
   

   

2. Can you tell me on which line of the code you get the 403 error? Is it during the process of uploading the artifacts? This error should happen when you submit the study. Is that the case?

[prateekwahi]

Hi Mostapha !!

1. I’m currently subscribed to the basic compute package, with an educational discount. and Yeah I just saw the runs per study limit, which I was not aware about.

2. I get error when it is done uplaoding all 1300 models. Maybe it is because of the 480 runs limit per study.

[mostapha]

Hi @prateekwahi,

Thank you for the quick response. That should be the reason. The server is doing what it is supposed to do. I assumed we also return a message that clarifies why you get the 403 error.

I’ll send you a private message to discuss the options to help you running your studies.

[prateekwahi]

Thankyou so much for your help !!