[Help] Power BI Report Slow on Initial Load (Live Connection)

[Zhinee]

Hi everyone,

I'm noticing an issue with one of my Power BI reports that uses a live connection.

• The initial load of the report is slow — it takes around 5–10 seconds.

• After that first load, the report becomes very fast.

• I've tried refreshing the page, clearing the browser cache, and even restarting my browser (Microsoft Edge), but after the first load, it always runs smoothly.

• However, if I come back after a few hours or the next day, the initial load is slow again.

I'm having trouble reproducing this consistently, but it seems like some kind of caching or timeout is involved.

Questions:

• What could be causing this "first-load slowness"?

• Are there specific caching mechanisms (on the browser, Power BI service, or the live data source) that I should investigate?

• What's the best way to debug this kind of issue?

Thanks in advance for any insights!

[DataVitalizer · Accepted solution]

Hi @Zhinee 
That “first-load slowness” is pretty common with Power BI reports using live connections, usually, it’s caused by the data source or Power BI service “warming up” caches or query plans after inactivity.

Here are few workarounds:

• Schedule periodic “keep-alive” queries to prevent cache timeouts.
• Optimize your data model and queries for faster execution.
• If possible, use Import mode for critical reports needing instant load.

Did it work? 👍 A kudos would be appreciated
🟨 Mark it as a solution to help spread knowledge 💡

🟩 Follow me on LinkedIn

[v-sshirivolu]

Hi @Zhinee ,

I wanted to check if you had the opportunity to review the information provided by @DataVitalizer . Please feel free to contact us if you have any further questions.

Thank you and continue using Microsoft Fabric Community Forum.

[Aburar_123]

Hi @Zhinee ,

you can try by enabling the below options in service to improve the performance,

[Zhinee]

I don't see that option on my end

[MasonMA]

Hi @Zhinee 

This is a feature only available for premium workspace. 

[MasonMA]

Hello @Zhinee 

I'd recommend going through below post from Community, which contains valuable insights to your question. 

https://community.fabric.microsoft.com/t5/Service/Slow-initial-load/m-p/1561975

As for optimizing report, the article “Troubleshoot report performance in Power BI” on Microsoft Learn describes how to investigate slow performance in Power BI, including behaviors in Premium or Fabric capacities. It outlines steps such as using Performance Analyzer, monitoring capacity metrics via the Fabric Capacity Metrics app, and distinguishing between types of slowdowns, such as dataset load vs. visual rendering. 

https://learn.microsoft.com/en-us/power-bi/guidance/power-bi-optimization

Hope this helps:) 

[DataVitalizer · Accepted solution]

Hi @Zhinee 
That “first-load slowness” is pretty common with Power BI reports using live connections, usually, it’s caused by the data source or Power BI service “warming up” caches or query plans after inactivity.

Here are few workarounds:

• Schedule periodic “keep-alive” queries to prevent cache timeouts.
• Optimize your data model and queries for faster execution.
• If possible, use Import mode for critical reports needing instant load.

Did it work? 👍 A kudos would be appreciated
🟨 Mark it as a solution to help spread knowledge 💡

🟩 Follow me on LinkedIn

[Zhinee]

Could you elaborate a bit on how to do this?

• Schedule periodic “keep-alive” queries to prevent cache timeouts.

[MasonMA · Accepted solution]

One of the ways would be use Power Automate to trigger a dataset refresh or run a query against a semantic model using Power BI REST API every 30 minutes.

However, Microsoft does not provide official documentation specifically endorsing the practice of keep-alive queries in shared non-Premium capacity workspaces. The concept of manually initializing datasets is more of a community best-practice workaround, rather than an officially documented feature.

[Zhinee]

I did try Run a query against a dataset and my query is :
EVALUATE
TOPN(1, 'SESSION_TRANS_TABLE')
and I set it run it every 30 mins, it still does not work.
Still slow on the initial load
Is there something wrong with the query or I should try to use refresh a dataset?

[MasonMA]

Hello @Zhinee 

I'm assuming you’re in shared capacity, you may also try changing the query to a more complex one such as join or summarize, 

Or use 

EVALUATE
TOPN(1000, 'SESSION_TRANS_TABLE'),

and Monitor whether first-load performance improves after scheduling it every 15–30 mins.

But again as I mentioned in my last message, Microsoft has no official document endorsing this approach. Shared capacity can also evict datasets even shortly after refresh, so no guarantee.