You’re offline. This is a read only version of the page.
401
Dataflows based on Power Query are a great tool for business users to manipulate data from multiple sources, thus automating otherwise tedious manual work processes. However, I am still weary to recommend using Power Query dataflows at scale due to inconsistent performance and server stability.
For internal reasons, our environments (and thus Power Query servers) are in the Europe region. I was wondering if users in other reasons were regularly (weekly) experiencing such issues:
1. We encountered an unexpected error
This cryptic message pops up way too often when using the editor, or trying to open the Power Query editor. And then it disappears, minutes or hours later.
2. There was a problem refreshing the dataflow. please try again later.
A problem, what problem? Later, how much later?
3. Dataflows that regularly take only a couple minutes to run occasionally take much longer (sometimes several hours)
The above considers very similar data volumes and identical, Microsoft cloud-based data sources(ie Sharepoint cloud and Dataverse via OData)
In any case, these messages should be more detailed.
Don't get me wrong, I absolutely love Power Query. I would just like to be able to rely on Power Query dataflows as much as I rely on other Microsoft products. As a Microsoft advocate within my organization, it is a little frustrating to have to go to my boss and say, "sorry, the dataflow crashed today, the numbers will be available later", "it crashed again? later, how long?", "later."
Thank you @bezhan-msft for your great work on this product. I am confident that the product will become more and more stable and will eventually be used even more broadly in the Power Platform (shameless plug, see my Idea post here : Action to transform data with Power Query - Power Platform Community (microsoft.com) )
Good to see that I am not the only one experiencing these issues with dataflows. My refresh failures fall into two categories:
I have documented these with my Power BI admin and even with our premium monitoring app in Power BI, we are unable to find a root cause. It is extremely frustrating as dataflows have huge value for usage in our organization but I cannot in good conscience recommend these to others with the current level of instability. Note, these large scale failures seem to only occur once our Power BI environment moved into Gen 2.
Hi! @CharlesS !! sorry for the late answer... the workaround i found was create a new environment and the dataflows run as normal, without problem...
Also i figured out to work better with solutions to carry to an environment from another... 🙂
Hello @Romansx22 !
I wish I could answer your question! Unfortunately, dataflows sometimes just fail for unknown reasons. Is your dataflow failing consistently, or was this a one-off problem?
Keep 'em flowin!
Charles
Hello @arpost !
Indeed, dataflows are great... when they work. To be used at an enterprise-level scale, performance stability and consistency is key. Unfortunately, dataflows aren't there yet. I feel like dataflows are currently akin to a preview feature of the Power Platform.
In my experience, most problems seem to coincide with back-end updates. One of my flows had been running fine for a few weeks up until Wednesday this week. That morning, I noticed that some icons appeared in the dataflow list page (the green check marks and red error icons were not there before):
Coincidentally, one of my V2 dataflows has been failing consistently ever since. (Note: Updates are deployed in waves across the world. While the icons appeared for me on Wednesday, they might have appeared to you before or might appear later). I have some serious doubts about the quality of the testing performed by Microsoft on new updates before they are pushed into production (specifically for Power Query dataflows).
I have an older, quasi-identical V1 dataflow as a backup that performs the same function. That dataflow is running fine. It is always frustrating when the newer versions of products are unable to perform the same tasks as the older, deprecated versions of the product.
I continue to believe that dataflows, and Power Query in general, have great potential. Actually, I would like to see their use expanded to other portions of the Power Platform, such as full integration with Power Automate (to perform data transformations within flows). Vote on this idea: Action to transform data with Power Query
I'm not convinced that Microsoft fully understands just how erratic dataflow performance actually is... and how important it is to invest time and ressources into ensuring performance stability and consistency. My advice: submit official support tickets. It is important for Microsoft to have documented evidence of all the problems customers are facing. Also, post on these forums regularly, with some luck we may get the right people's attention.
I have faith dataflows will eventually be stable enough to be used enterprise-scale. For now, let's keep on pushing Microsoft to improve performance consistency and stability!
Keep 'em flowin!
Charles
Right there with you, @CharlesS. I’m in the US region and find performance and throughput to severely limit Dataflow/Power Query adoption. I have small files (2 rows) that have taken around 5 minutes to process, and my UI keeps having to be refreshed randomly while other sites work just fine.
I’m a major Power Platform advocate and fully embrace the value of low-code/no-code; I use Power Query every chance I get where I can. Sadly, I find that we still can’t really offer Dataflows as a solid alternative to pro-code approaches to enterprise-level, non-analytical ETL.
Don’t get me wrong; I love Dataflows and value the work Microsoft has done on them; I just can’t recommend them (or even really use them) to achieve large-scale tasks until the limits/performance gets worked out. Well, that and the ability to pass parameters to Dataflows from outside the dataflow…but that gets into the feature request I’ve submitted. 😛
#1
mmbr1606
9
Super User 2025 Season 1
#2
stampcoin
7
#3
SD-13050734-0
6
Share
Report
All responses (
Answers (
