Direct Query dashboards are too slow and widgets fail in QuickSight

Q&A
,
1

Hello Amazon QuickSight Community,

We are using Amazon QuickSight through Siemens Opcenter Intelligence for Cloud and we are experiencing performance and reliability issues with Direct Query datasets.

When users open a dashboard or interact with filters, some visuals/widgets take too long to load. In several cases, the widgets eventually fail and show an error instead of displaying the data.

We tested the same scenario using SPICE and the performance was much better. However, Siemens release notes for Opcenter Intelligence for Cloud state that SPICE capacity allocation cannot be guaranteed for cloud users, and therefore Siemens strongly recommends using Direct Query mode. Because of this, we are trying to optimize the Direct Query approach.

Environment:

  • Amazon QuickSight through Siemens Opcenter Intelligence for Cloud
  • Dataset mode: Direct Query
  • Issue: Slow dashboard loading and widget/visual errors
  • SPICE result: Better performance, but not recommended/guaranteed for this cloud environment

We would appreciate guidance on the following:

  1. What are the best practices to improve Direct Query performance in QuickSight?
  2. How can we identify which visual or generated query is causing the delay or timeout?
  3. Are there known limits or timeout conditions that can cause visuals to fail in Direct Query mode?
  4. Is it recommended to move calculated fields and transformations to the source/database layer?
  5. Would using aggregated views, pre-filtered datasets, or database-side optimization help?
  6. Are there dashboard design recommendations for Direct Query, such as limiting the number of visuals, filters, calculated fields, or high-cardinality fields?
  7. Is there a way to monitor or inspect the SQL queries generated by QuickSight for each visual?

Any recommendations, troubleshooting steps, or best practices for using Direct Query in this type of setup would be appreciated.

Thank you.

image
image1163×264 28.7 KB

Thank you.

2

Hi @agustinmoncada94,

Nice seeing you back in the Quick Community! To improve your Direct Query performance, I would recommend checking out metrics of your visuals within CloudWatch, as that will give you a better idea of which specific visuals are taking longer to load (or causing timeouts). Here is some documentation that can guide you in this process.

Let us know if this aids your use case at all. If not, it definitely would not hurt to test bringing any calculated fields/parameters over to the dataset prep level (or explore other workarounds in general).

Thank you!

3

Hi @agustinmoncada94,

Just checking back in since this thread hasn’t received a response in a while. Was my most recent reply helpful to you and/or were you able to find a solution yourself in the meantime? Please help the community by marking this answer as “Solution” or following up in general within the next 3 business days!

Thank you!

4

Hi @agustinmoncada94,

Since I haven’t received any further updates from you, I’ll treat this inquiry as complete for now. If you have any additional questions, feel free to create a new post in the community and link this discussion for context.

Thank you!