How do I write a monitor for an app?

☁️ Community Cloud
1

Hi,

I’ve got an app that we advertise broadly on the streamlit sharing platform:
https://share.streamlit.io/jkanner/streamlit-dataview/app.py

It works well most of the time, but it occasionally crashes - maybe once every couple of weeks. While I’d like to make the app more stable, I’m not sure how to debug this.

My question is: Is there a way to write a monitor for streamlit apps? For example, if I access the app URL with an automated GET request, is it possible to tell from the response if the app is running? I can live with the occasional crashes, but I’d love to get alerted when it needs re-started, or even better, have the app re-start automatically after a crash.

I’d appreciate any hints or tricks. Thanks for considering!

Best,
jonah

1 Like
2

Yes, of course.
There would be quite a few different possibilities.
There are hosted health check services.
You could host a health checker yourself.
You could program a health checker yourself.
If yes, on which platform and in which programming language?

3

@Franky1

Thank you for the quick response!

Well - I’m open to any suggestions you have for health checker’s. I don’t have much experience with these.

I am working on a linux machine, and I don’t mind writing my own monitor. I’m most comfortable working in python, but I could use bash or some other language if needed. There are also members of my team who use nagios monitors. I could try to plug-in to that infrastructure, though I’d have to get myself up the learning curve.

In any of these scenarios, I think I’d need to know what to look for in an HTTP response to tell if the app is running (right?).

Again, thank you for your help!

4

So if you already have a running monitoring solution in your organisation, I would definitely use that and not reinvent the wheel. With Nagios, that’s definitely possible, but I’m not familiar with it, so you’ll have to ask your colleagues. :wink: