How to investigate Quicksight Dashboard Load time spikes

Question & Answer
,
1

We have set up alarms for dashboard load time on one of our Quicksight Dashboard (pivot tables) that consumes two SPICE datasets (one data set has about 1100 rows, the other has ~14000 rows). In general our dashboards take about 3-5 seconds to load, but we saw a couple spikes to 8 seconds.

What are some ways I can root cause this latency issue/ what are some potential causes for spikes in latency for a QS dashboard? Any guidance would be much appreciated!

2

Hi @dting

Welcome to the community!

To root cause the latency issue in your Amazon QuickSight dashboard, here are some steps

  • SPICE Usage: Check if your SPICE capacity is reaching its limits. High usage or nearing full capacity can slow down queries and increase latency.

  • Data Refresh Schedule: Ensure that data refreshes or ingestions are not happening during peak usage times, which could lead to delays

  • Visualization Complexity: Pivot tables with multiple measures, dimensions, or filters can increase load times. Try simplifying the pivot tables to see if that improves performance.

  • Calculated Fields: Check if you aren’t running a large number of calculated fields in your analysis. Calculated fields can add significant overhead, especially if they are complex or numerous.

  • Direct query datasets can cause significant delays since they retrieve data in real-time from the source.

If your analysis/dashboard doesn’t have many calculated fields, I suggest reaching out to AWS Support for further assistance. They can look into the specifics to provide more help. Here are the steps to open a support case.

1 Like
3

Could you talk a bit more on visualization complexity? Is there documentation I can read into about how different measures/dimensions/filters affect the latency of pivot table load time?

And is there any way to analyze spikes of latency? The points mentioned above are great for reducing latency, but I’d like to know more into why a particular loading of a pivot table took longer compared to a different loading of the same table.

Thanks!

4

Hi @dting

When dealing with visualization complexity, several factors can affect the latency of loading pivot tables, including:

  1. Measures and Dimensions: The number and complexity of measures and dimensions directly impact the processing time. Complex calculations, multiple measures, and high cardinality dimensions (dimensions with many unique values) can increase the load time as QuickSight needs to process more data.
  2. Filters and Slicing: Applying filters, especially those that are computationally intensive (like filters on calculated fields), can slow down loading times. The more filters and conditions you apply, the more processing QuickSight has to do, which can increase latency.
  3. Visual Complexity: The complexity of the visual itself—such as the number of rows and columns in a pivot table, or the use of conditional formatting and calculated fields—can affect performance. More complex visuals require more processing power and time to render.

Please refer the below documentation this might be helpful for you.

1 Like
5

Thanks! Is there any way to investigate a specific loading instance that took longer than usual? For example, we have a dashboard pivot table that normally takes about 2 seconds to load, but for one instance it took 8 seconds to load. Are there any tools I can use to debug this specific instance (i.e cloudwatch)?

6

Hi @dting,
I do not believe that there is currently a way to investigate the loading performance of visuals in QuickSight.

Did you have any additional questions regarding your initial topic? If we do not hear back within the next 3 business days, I’ll go ahead and close out this topic.

Thank you!

7

Hi @dting,
Since we haven’t heard back, I’ll go ahead and close out this topic. However, if you have any additional questions, feel free to create a new topic in the community and link this discussion for any relevant information that may be needed.

Thank you!