[Python] Is there a way to achieve near-instant updates in Streamlit using GridDB as a...

I'm working on a real-time monitoring dashboard using Streamlit, with GridDB as the backend time-series database. My goal is to update the dashboard with new data every second.

Here is my current setup:

I'm using the griddb_python_client to connect and query GridDB. The dashboard refreshes periodically using st.experimental_rerun() to display the most recent data.

However, I'm running into a limitation:

GridDB does not provide a native pub/sub mechanism or event triggers to notify the application when new data is inserted. This forces me to rely on polling, which is not efficient.

Could there be a smarter or more efficient way to integrate GridDB with Streamlit for near-instant or real-time updates — ideally without resorting to constant polling?

Here is a snippet of my code

def get_latest_data():
query = container.query("SELECT * WHERE timestamp > TIMESTAMPADD(SECOND, -5, NOW())")
rs = query.fetch()
return [rs.get_next() for _ in range(rs.size())]

st.title("Real-time Dashboard")

data = get_latest_data()

st.write(data)

time.sleep(1)
st.experimental_rerun()

Continue reading...