Link to app: https://footverse.streamlit.app/
Hi everyone — I’m experiencing frequent crashes on my Streamlit Cloud app with the generic error message:
“Oh no. Error running app. If this keeps happening, contact support.”
Context
- The app loads a DataFrame from a database and uses
@st.cache_data
to cache it. - DataFrame shape: (8369, 315).
- I load filtered chunks per page from the cached function (not all at once).
- I store user filter selections in
st.session_state
. - Previously, when the DataFrame was smaller (e.g., 2500 rows), the app worked fine — even when I was (mistakenly) storing it in session state.
- The app has recently seen increased traffic — ~200 users/day.
- Now the app keeps crashing sporadically.
What I’ve Tried
- I removed storing the DataFrame in
st.session_state
. - Used
memory_profiler
locally — peak memory usage ~650MB. - The app works reliably in local testing with the full dataset.
- No detailed error messages or logs are shown in the cloud logs — just the generic crash error.
Environment
- Streamlit version: 1.45.0
- Python version: 3.12
- Requirements:
streamlit
pandas
plotly.express
plotly
lxml
matplotlib
streamlit-javascript
scikit-learn
umap-learn
firebase-admin
orjson
gdown
# webdriver-manager
# undetected-chromedriver
# selenium-stealth
Looking For
Any ideas or suggestions on:
- Whether the issue is memory-related on the Streamlit Community Cloud backend?
- Ways to optimize or paginate better?
- How to get better crash logs or isolate what’s failing?
Thanks so much in advance