The service has encountered an error while checking the health of the Streamlit app: Get "http://localhost:8501/script-health-check": EOF

[Monica_Molinas]

Hi! I am trying to deploy on streamlit. However I have the following error:

[12:20:42] Processed dependencies!

[12:20:55] ❗️ The service has encountered an error while checking the health of the Streamlit app: Get "http://localhost:8501/script-health-check": EOF

[12:22:27] Streamlit server consistently failed status checks

[12:22:27] ❗️ Please fix the errors, push an update to the git repo, or reboot the app.

Here is my git: GitHub - nuggi46/streamlit
Here is my app: https://foodies.streamlit.app/

Please help me!

Thank you

[ferdy]

Does your app work locally?

[Monica_Molinas]

Yes, my app works locally!

[ferdy]

Looks like the max memory allowed has been exceeded.

# df_sample = df_restaurante.head(20000)
df_sample = df_restaurante.head(5000)

In my computer, if I use 5000 it is already using 1GB of RAM. 1GB is the limit in Streamlit community Cloud.

Try to use say 1000.

Or even better, play around with the app on your PC and record its highest memory usage. Then quit the app. Then measure the memory difference. If its too high reduce the sample size. Then deploy it in the cloud.

Also try to cache your data.

@st.cache_data
def get_restaurante(fn):
    return pd.read_parquet(fn)

ruta1 = "https://storage.googleapis.com/yelp-and-maps-data-processed/df_resto_user_final.parquet"
# df_restaurante = pd.read_parquet(ruta1)
df_restaurante = get_restaurante(ruta1)

[Goyo]

user_resto_matrix has 20000 rows and 890 columns that need 136.0 MiB of storage. In my tests, calling cosine_similarity on it seemed to exhaust the resources available in streamlit cloud. Note that the output would be a 20000 x 20000 array, more than 3GiB if my maths are correct.

[Monica_Molinas]

Thank you! It works!

[Monica_Molinas]

Thank you! Now I did with a sample of 1000 rows and the app is working!