How to run Dash App on Azure web services asynchronously for multiple users?

Dash Python
1

Hi, I have two questions:

I created a Dash chatbot that takes a query from users and returns the most relevant documents from the database using SentenceTransformers, FAISS, DistilBert, etc. I deployed the app using Azure Container Registry and App Services. When different users reach at the website, one can see others interaction with the chatbot and I cannot figure out if this is an issue with Dash App or the App Services.

Currently, I am running the app on a single container on Azure that takes the code/data from Azure Container registry instance I’ve created.

I tried to add threaded=True or processes > 1 within app.run_server( ) function but they didn’t help either.

  1. How can I make this app available to different users asynchronously?

  2. I am also trying to get user feedback after each conversation and store that data to Azure / GitHub somehow, and each time a new conversation starts it will need to load the data and append the new feedback data & write it to Azure storage / GitHub. Is this achievable with Plotly Dash?

Thanks in advance!

2

Are you using global variables?

3

Yes, some of them are global. Is that the issue?

4

Yes. If you remove the global variables, your app should work as intended.

1 Like
5

Let me try. Thank you very much!

Another question: If I want to use a variable within multiple callbacks is there any other way than declaring them global? I think that was the main reason why we used global variables in the first place.

6

If the data are not too large, you can use a Store component.

7

See Part 4. Sharing Data Between Callbacks | Dash for Python Documentation | Plotly

8

Thank you! Let me look.