permissions issue with Sagemaker endpoint deployment

Question

Hi there,   
  
I am having permission issue deploying a SageMaker Endpoint, if someone could help me out here.  
  
If I run this code in Sagemaker's Jupyter Notebook without providing credentials for boto3 client, it works. However if I try to invoke it with my credentials on my local machine, it gives me permission error, even though my account already was given all Sagemaker permissions.   
  
**ClientError: An error occurred (AccessDeniedException) when calling the InvokeEndpoint operation: User: arn:aws:iam::XXXXXXXXXXXX:user/jenny.lien is not authorized to perform: sagemaker:InvokeEndpoint on resource: arn:aws:sagemaker:us-east-1:249707424405:endpoint/XXXXXXXX with an explicit deny**  
  
This stackoverflow post has exactly the same issue as what I've experienced, I tried disable MFA for my IAM user but it is still not working...does it mean I will also have to create a special IAM  user to make it work? thank you  
https://stackoverflow.com/questions/54172907/amazon-sagemaker-accessdeniedexception-when-calling-the-invokeendpoint-operatio

adambookletai · Answer

Hey there,   
  
Odd that you are still running into this issue after turning off MFA. I usually debug issues with IAM by creating a new user that is cleared from all other settings. Have you tried that approach yet?  
  
Also, if you're looking for a no-code option to create an HTTP API for a Sagemaker endpoint, checkout https://Booklet.ai. If you point Booklet.ai to your endpoint, it will handle these invoke calls. From there, it will wrap your model endpoint in an HTTP API, provide a responsive web demo form, and more. You can bypass this invoke function.  
  
Hope this helps!

awsjennyaws · Answer

just wanted to update that I didn't turn of MFA the proper way. After doing so it works now. Thanks!

permissions issue with Sagemaker endpoint deployment

Relevant content