Dataflow Refresh Failing

[efh3]

Hi,

I have a Dataflow that is scheduled to refresh twice a day. Over the weekend, it successfully refreshed a couple of times, but then started to fail. I do not get meaningful error messages. Instead, when I download the CSV file for the refresh, it says thing like "Error: Internal error Request ID: 1a74efbc-da2d-d0ef-707f-0efd86741928 Activity ID: 373515be-c301-4a3e-969e-d2b52145e3b0".

I have also received occassional timeout errors recently, which may or may not be related, since the most recent errors do not mention timeouts, but they started at around the same time.

If I edit the dataflow and try to preview the data, I do not seem to get any errors with the subset of data that the system ingests under such conditions.

Is there any way for me to obtain any more meaningful logs for debugging purposes?

Best

[efh3]

Just to be clear, between the last time the automated refresh worked and its first failures, both over the weekend, I made no explicit changes to the Dataflow code, nor any of the applications to which it connects, nor to any permissions of which I am aware. Now, every time I try to manually run a refresh, it fails, but if I edit the Dataflow and refresh a preview, it works.

TY for any suggestions on how to debug deeper.

[efh3]

I am now seeing errors from Power BI Service including

"us.prod.powerquery.microsoft.com took too long to respond"

when I first open up my Dataflow to edit, or, if it gets past this, messages about unknown error when trying to present me with a mashup - unfortunately, I closed the browser and cannot reproduce that error anymore, since Power Query just shows the spinning "working on it" swirl over an empty thought bubble icon.

[apotter6]

Experiencing the same issues.  

Azure SQL and Sharepoint (cloud) data sources for the flows.  

Always dies after 1:25 

[Anonymous]

We had the same issue where we used to get internal service error and refresh was failing fairly regularly,  what we did was we manully refreshed quite a few reports at the same time and looked at the Power BI server and found that the RAM was going over the limit, CPU was just using 20 to 30 % but RAM was 100% when a lot reports are refreshing at the same time.

Hardware team increased the RAM and all our reports are being refreshed fine.  Hope this helps you.

Nick.

[efh3]

TY, all, for your feedback and insights.

My queston is more about getting better debugging information than solving the root problem, per se.

In this particular case, I suspect that the problem was with the Dataflow's connection to one of its data sources - but the error message provided doesn't really help me with that diagnostic.

FYI, I am pulling data from a variety of sources, including JIRA. It would appear that the app team that own one specific JIRA filter made a change such that our ETL Service Account could no longer execute the filter. My challenge occurs when changes are made to these JIRA filters without me being told - Filters for which my ETL is dependent but over which I have limited control or awareness (since I do not own these JIRA filters),

• How can I get more meaningful errors than simply something like
  • "Error: Internal error Request ID: .... Activity ID: ...".

This error gives me almost no insight into what the real problem is.

• Are there any debug flags that I can set for the Dataflow?

Unfortunately, we do not have a Premium license. If we did, I would surely be looking into creating deployment pipelines, which might provide more feedback or at least alternative ways to break up the ETL and possibly localize the source of the problem more quickly.

Thanks again - and good luck with your own related issues!