Forum Discussion

Hussein_charif's avatar
9 months ago
Solved

Report performance on fabric workspace

Hello all,

i have a report published on a fabric workspace. the capacity is F16, we have around 15 users and growing, the capacity usage average is between 50-70% daily. the report is directquery reading from a warehouse view that updates every minute.on power bi desktop, it is very fast in terms of performance, but on the service, it takes around 5-10 seconds to load sometimes. anyone has an idea what could be the issue?

 

also to add, sometimes, when we first open the report on service, there are missing values/numbers, we'd have to refresh the page or reset filters to get the missing values. what can cause this problem?

5 Replies

  • Hi Hussein_charif, 

     

    This is expected as the service needs to wait for a query to return from the datasource before it can render a visual, and in the service there is more network latency as your datasources move through on prem gateways or through different datacenters around the world. 

     

    If you want to avoid the delay, I'd recommend moving towards using a Lakehouse or Warehouse to store your data, and using a DirectLake storage mode. 

     

    Direct Lake overview - Microsoft Fabric | Microsoft Learn

  • The issues you see are usually caused by a combination of capacity pressure, DirectQuery concurrency, view complexity, and query cold-start effects.

    The “missing values” issue is often related to partial query execution, visual query cancellation, or timing issues with DirectQuery results, not data corruption.

     

    You can replace the view with a materialized table, Reduce visuals per page if possible, and Check Fabric Capacity Metrics app for impact.

     

    References that can help you : 

  • Hello Hussein_charif 

     

    Reason 1: Warehouse Performance & Query Overhead

    • When you use DirectQuery, all the SQL runs right in your warehouse. If the queries aren’t tuned up, things can slow down a lot.
    • If you’ve got more than 15 people making frequent updates, the database gets hit pretty hard.

    What you can do:

    • Fire up Performance Analyzer to spot heavy queries.
    • Add indexes, partitions, or use materialized views to pre-crunch the data and lighten the load.
    • Stick to Fabric best practices—like using V-Order, keeping the number of parquet files low, and optimizing delta tables.

    Reason 2: Latency from Live Queries & Capacity

    • Every visual sends off its own query, so you can get sudden usage spikes.
    • Even if your F16 is running at 50–70% most of the time, lots of live queries can still slow things down.

    How to handle it:

    • You might want to bump up capacity or scale out to avoid throttling.
    • Keep an eye on peak usage and query throttling in the Fabric Monitoring app.

    Hope this helps - please appreciate by leaving a Kudos or accepting as a Solution! 

  • Thankyou, AntoineW, tayloramy and deborshi_nag for your responses.

    Hi Hussein_charif,

    We appreciate your inquiry through the Microsoft Fabric Community Forum.

    We would like to inquire whether have you got the chance to check the solutions provided by AntoineW, tayloramy and deborshi_nagto resolve the issue. We hope the information provided helps to clear the query. Should you have any further queries, kindly feel free to contact the Microsoft Fabric community.

    Thank you.

  • Hi Hussein_charif,

    We are following up to see if what we shared solved your issue. If you need more support, please reach out to the Microsoft Fabric community.

    Thank you.