After reviewing observability data, you find that Tableau Server’s data extract refreshes are significantly impacting performance during business…

The problem states that extract refreshes are impacting performance specifically "during business hours." This indicates a resource contention issue where background tasks (refreshes) are competing with interactive user sessions for CPU, memory, and I/O. The most direct and architecturally sound solution is to manage this workload by scheduling the resource-intensive extract refreshes to run during off-peak hours (e.g., overnight). This separates the workloads, ensuring that server resources are dedicated to user-facing activities during the day, thereby resolving the performance degradation without altering the data strategy or incurring unnecessary hardware costs.

A. Moving all data extracts to live connections is a drastic change that shifts the performance burden to the underlying database and may not be optimal for complex dashboards.

C. Completely disabling extract refreshes is not a viable solution as it would result in stale, outdated data, rendering the analytics useless for decision-making.

D. Upgrading the server's CPU is a reactive, "brute-force" approach. While it might shorten refresh times, it doesn't solve the fundamental problem of resource contention during peak hours.

1. Tableau Server Documentation: In the "Performance Tuning" section

a key recommendation is workload management. It states

"Schedule extract refreshes for off-peak hours. Refreshes are resource-intensive jobs. Therefore

it is best to run them outside of business hours."

Source: Tableau Help

"Performance Tuning

" Section: "Tune for Extracts."

2. Tableau Whitepaper: The official "Best Practices for a Performant and Stable Tableau Server Environment" whitepaper explicitly advises on this scenario. It highlights the role of the Backgrounder process

which handles extract refreshes

and recommends scheduling these tasks to avoid conflict with user activity.

Source: Tableau Whitepaper

"Tableau Server Scalability

" Section: "Workload Management

" Page 12.

3. University Courseware: Materials on Business Intelligence systems architecture from institutions like the University of Washington explain the different processes within Tableau Server. They detail how the Backgrounder process handles asynchronous jobs like extract refreshes

while the VizQL Server process handles interactive user requests. To optimize performance

these competing workloads should be segregated by time.

Source: University of Washington

"Tableau Server Architecture & Administration

" Section: "Tableau Server Processes

" Slide on "Backgrounder."