PowerBI service correctly running ARIMA forecasting, but not SARIMA

[dogsPlayingPokr]

I have two different reports, one forecasting against test data using an ARIMA model, the other using a SARIMA model against the same data.
Both work correctly in local, and as per Learn which Python packages are supported - Power BI | Microsoft Learn  both are called under the statsmodel package which is supported. All other packages used are also among the supported list. 
When published to service, the one running on the ARIMA model is correctly functioning, but my SARIMA model is not replicating itself on the service. 

Examples: Correctly functioning SARIMA prediction when working on desktop. 

-----------------------------------------------------------

The same model when published to the service.

My ARIMA model works the same both on the service and locally: Is there some difference in execution that I need to accomodate for?
If I were using an unsupported package, my script would fail to run completely if published. 

[Anonymous]

Hi, @dogsPlayingPokr 

Make sure that the Python environment on your local machine matches the environment on the Power BI service. Sometimes, differences in package versions can cause discrepancies. Double-check the version of statsmodels and any other dependencies.
The Power BI service might have resource constraints that affect the execution of more complex models, such as SARIMA. Try simplifying the model or reducing the data size to see if it makes a difference
Check the error log in the Power BI service to see if there are any specific error messages when the SARIMA model fails. This provides a deeper understanding of what may be going wrong. If the SARIMA model takes longer to execute, it may be a Power BI service timeout. Try optimizing your scripts to improve performance.

How to Get Your Question Answered Quickly

Best Regards

Yongkang Hua

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.