I have noticed a recurring issue with my Quicksight dashboards where, despite successful query executions and dataset refreshes, the dashboards occasionally display incorrect numbers and data deviations. While manually refreshing the dashboard resolves the issue and displays the correct data, I’m concerned about the root cause. I need help understanding why this inconsistency occurs and would like to implement a monitoring mechanism to prevent such discrepancies in the future.
Hi @javvajim
Based on your general comments it appears that you have an incremental dataset and if that is the case it is possible that you may be having changes in your underlying datasource but those are not being captured in your incremental dataset.
Have you looked at that angle? Or are there any other insights you have about the issue that you can share with us so we can comment further?
Regards,
Giri
Hi Prabhu,
My job is not an Incremental Refresh, Mine is a full refresh and it refreshes daily at once at 2:00PM IST.
Hi @javvajim - Please ensure the full refresh schedule should happen after your underline data refresh. If there is any overlap between data refresh and spice refresh, you will observe data deviations..
Regards - Sanjeeb
My underlying data sets are refreshed every day at 00:00 hr in data net in America/Los Angeles time zone where as my quickisght data set is refreshed at 2:00 PM IST time zone. So I don’t think there is an overlap between both the refreshes. Even if that is the case , I have a load job in the data net where the data gets stored in the table. So Ideally even though there is an overlap in the datarefresh, since the load job is not executed , ideally quicksight should fetch the data from the table of the previous days . Even the previous day numbers in this case are same and even then there is deviation in the data. So I don’t think so its a data refresh issue.
Can some tell me what is the issue here or is there any ticket that I can raise to the team regarding the same