Dashboard performance problem

Jones01
Jones01 🟑

Hi guys,

I have a view that once built has 112k rows. It is built from 7 datasets.

Am I right in thinking once the view is built then it doesn't need to go back to the datasets and perform joins etc?

I am seeing some very poor performance on the dashboard built from this view but the number of rows is quite low.

The widgets I am using are all period over period widgets.

Would you suggest anything obvious to check?

Thanks

Answers

  • amehdad
    amehdad 🟀

    Hi @Jones01, one suggestion is get your dataset, which powers your dashboard, to only have the relevant data for those views (as opposed to fetching all the possible records from those 7 inputs datasets). This can be done on the ETL v2, where you can filter records or remove columns, so that the processing time is reduced.

  • amehdad
    amehdad 🟀

    One another note, I made an idea suggestion around having the ability to pinpoint areas around dashboard performance: https://dojo.domo.com/main/discussion/53832/dashboard-performance-report. Please upvote if you agree

  • MichelleH
    MichelleH 🟒

    @Jones01 112k rows should not be enough to cause performance issues on the dashboard. More likely it would be caused by the number of cards, the granularity of the cards, and the internet connection/memory capabilities of your individual computer.

  • @Jones01 - I believe that dataset views are also not materialized datasets. Meaning that for each card on your dashboard, Domo is first performing the joins and then rendering the visual. If you are noticing performance issues, I might suggest joining the data in an ETL or SQL dataflow first. Alternatively, you could reach out to your Domo CSM to see if they can materialize the view for your instance to improve performance.


    ______________________________________________________________________________________________
    β€œThere is a superhero in all of us, we just need the courage to put on the cape.” -Superman
    ______________________________________________________________________________________________
  • Jones01
    Jones01 🟑

    @ST_-Superman-_ Thank you. I had been under the impression they were.

    I have also read about Adrenaline DataFlows. I thought what we had, used what ever Adrenaline is 🀣