Streamlit download_button while downloading the excel file its showing error file was not available on the site on first 4 download clicks

Using Streamlit
,
1

I have a python code to create an excel file with multiple sheets from data frame.
The below is working as expected in local but when the code is deployed to beta and production. The initial 2 to 4 clicks are not downloading the file into machine later its downloading the file content.

Code
@st.cache_data(ttl=300)
def create_text_preference_excel_data(df1, df2, df3, df4, df5, df6):
“”"
Save multiple DataFrames to an Excel file with separate sheets.
“”"
output = io.BytesIO()
with pd.ExcelWriter(output, engine=‘xlsxwriter’, engine_kwargs={‘options’: {‘in_memory’: True}}) as writer:
df1.to_excel(writer, sheet_name=‘WorkItem Level Feedback’, index=False)
df2.to_excel(writer, sheet_name=‘Summary’, index=False, header=False, startrow=0, startcol=0)
df3.to_excel(writer, sheet_name=‘Summary’, index=False, header=False,
startrow=len(df2) + 3, startcol=0)
df4.to_excel(writer, sheet_name=‘Summary’, index=False, header=True, startrow=len(df2) + len(df3) + 4, startcol=0)
df5.to_excel(writer, sheet_name=‘Summary’, index=False, header=True, startrow=len(df2) + len(df3) + len(df4) + 5, startcol=0)
df6.to_excel(writer, sheet_name=‘Summary’, index=False, header=True, startrow=len(df2) + len(df3) + len(df4) + len(df5) + 7, startcol=0)
return output.getvalue()

excel_data = create_text_preference_excel_data(df1, df2, df3, df4, df5, df6)

left, right = st.columns([0.87, 0.13])
with right:
    st.download_button(
        label="**Export Report**",
        data=excel_data,
        file_name="report.xlsx",
        mime="application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
        disabled=st.session_state.disabled_buttons
    )

Error Message
In Chrome Browser - File wasn’t available on the site
In Firefox Browser - File Not found error

Versions
Streamlit Version - 1.31.0
Python Version - 3.9

The same code works with 100% reliability locally but once its deployed via aws cloud its not downloading the file in initial clicks and later after some time its downloading the file.

Your help is highly appreciated.

Thanks.

2

Pretty sure this is because in the cloud, there are multiple replicas, and each request can go to any of the replicas. So, the request to get the file is hitting the replica that doesn’t have the file. I thought I configured my app with “sticky routing” to fix this but I still sometimes get this error so I’m not sure what the solution is.