Consider AWS services to deploy Streamlit application

Using Streamlit
,
1

I’m searching for some beneficial advices from the Streamlit community on the best way to deploy my Streamlit application on AWS. The app is a chatbot that interacts with Amazon Bedrock Knowledge Base.

I’m considering the following AWS services for deployment:

** App Runner**
** AWS Fargate**
** Elastic Beanstalk**
** EC2**

When choosing a deployment option, I’m considering the following factors:

  1. Auto scalability
  2. Request and response time
  3. Support for concurrent 1000 users
  4. Reliability

Here’s an overview of my application:strong text

  • It consists of several Python files:
    • app.py: Renders Streamlit user components like chatbot message controls and select boxes.
    • bedrock_kb.py: Interacts with Amazon Bedrock service using boto3 and AWS Glue connection with awswrangler.postgresql.
    • util.py: Utility file.
    • login_ui.py: Handles Azure AD SSO using the Microsoft Authentication Library (MSAL) for Python.
    • .streamlit/secrets.toml: Stores credentials for Azure AD SSO.
    • env: Stores other configuration settings.

Can anyone recommend the most appropriate, simplest, and cost-effective way to deploy my Streamlit application on AWS, considering these factors and the specific requirements of my app?

Thanks in advance for your advice!
Pongthorn Sangka

2

This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.