We’ve built several ad-hoc dashboards in Tableau Desktop 2023.1 for real-time monitoring of our operations data. The dashboards use live connections to our streaming database via ODBC connector, but we’re experiencing significant refresh delays - sometimes 45-60 seconds between updates even though the underlying data changes every 5-10 seconds.
I’ve tried adjusting the dashboard refresh interval settings in the workbook, but it doesn’t seem to make much difference. The live connection mode was chosen specifically to avoid extract refresh overhead, but now I’m wondering if that was the right architecture decision.
Has anyone dealt with ODBC streaming connector limitations in live mode? Should we be looking at extract mode with more frequent refreshes instead? The delays are impacting our ability to monitor real-time operations effectively.
The 45-60 second delay makes sense now. With multiple LOD calculations across sheets, Tableau is executing separate queries for each calculation, and ODBC connections serialize these queries. Each query might only take 2-3 seconds, but with 15-20 queries running sequentially through the ODBC layer, you get your minute-long refresh cycle. Live connections don’t benefit from Tableau’s query optimization the way extracts do.
Live connections can definitely struggle with streaming data sources. The refresh delay you’re seeing is often due to query execution time rather than the connection type itself. Have you checked the query performance on your database side? Sometimes the ODBC driver adds latency too, especially with complex calculations in your views.
Thanks for the suggestions. I checked our LOD calculations and we do have several across different sheets in the dashboard. The database query performance looks fine when tested directly - usually under 2 seconds. So it seems like the bottleneck is really in how Tableau handles the live connection refresh cycle with ODBC.