Spider-Sense | browser extension to monitor your spiders | Browser Plugin library

I'm a Data Science student finishing the thesis project and a Dash rookie.
I've created for my internship some analysis tools that I will have to share with many users inside the company, and by checking all of the possibilities to deploy the tools I decided to use Dash to let my co-workers work with the dashboard by using the browser. The dashboard is already functioning on my laptop, but I'm currently experiencing headhaches to understand how could I share it by using a Linux VM that my company provides me.
It's the first time that I have do deploy something i have created and i'm in trouble with it.

Basically, the dashboard does create some matplotlib plots based on some user input (thanks to dropdowns) and makes it available to the user by returning an encoded image of the plot (why? it was the easiest way for a Dash rookie). This image is saved by some "backend" functions in a local directory every time the function is called (by the callback function that i share later in the question), so I suppose that once I deploy the dashboard on the VM an awful event could happen: every time a n+1 user asks to the dashboard to create a plot, the functions that create the plot could overwrite the plot figure that a n user is trying to obtain. I don't know if this concurrency problem will actually happen when I'll deploy the dashboard onto the VM, but i would like to hear any possible solution to avoid this: my spider-senses are telling me that this will happen.

Just to make it clear before someone asks me, I decided to avoid using plotly because the functions necessary to create the matplotlib plots were already available and cannot be changed. So, as a Dash rookie, I found this escamotage to plot the data using images, but now I'm kinda regretting it. I'd like to hear any other possible solution as I didn't find anything helpful on the Dash documentation.

Considering the code on which the dashboard relies, there are some "backend" functions (function1, function2, ...) that I use to create different plot types based on different inputs. The plot is saved in the folder, and the following update_spec callback function calls the "backend" function and pick the result from the folder, sending it to the dashboard Img component to plot it as an image. The callback function (simplified) is something like: