Forum Discussion
Intermittent API flows failing to sync
Hi,
We recently updated our flows to include an API key due to a connected system depreciating basic auth.
Our flows have run fine before, and a few times since updating but now seem to be failing pretty regularly on the schedule (have worked manually since).
Any ideas?
Request ID: ccf751bf-c706-bedc-ef2a-ecbe9df461cf Activity ID: 08587c10-0e38-49ce-b6e9-bc125d36fce1
We don't get much more info in the error csv so wondering if someone can point me in the right direction of where to look next.
Time out, merge issue?
Thanks,
N.
10 Replies
Hi NiallFabric ,
Thank you for reaching out to Microsoft Fabric Community.
Thank you Poojara_D12 Ahmed-Elfeel for the prompt response.
I wanted to check if you had the opportunity to review the information provided and resolve the issue..?Please let us know if you need any further assistance.We are happy to help.
Thank you.
Could you please provide the error code From Power Automate portal > Flow > Run History > failed run > Download Input-Output.
it could be API rate limit or something else but first we need the error code before providing any solutions.Hi,
The flows sit within Power BI online (they don't appear in my Power Automate even though I am listed as the owner), when I go to the run history of the flow in Power BI the only detail I get is the code originally pasted beside "failed".
Error: Request ID: ccf751bf-c706-bedc-ef2a-ecbe9df461cf Activity ID: 08587c10-0e38-49ce-b6e9-bc125d36fce1
Is there away to get more detail on that?
Help articles suggest MIcrosoft can research these further, so might need to go to them?
Thanks,ok lets Fix this in Points:
1-First you should go to your Dataflow settings in PowerBi Service then open the connections tab (This is where you can see whats powering your refreshes behind the scenes)
2-Check the connection type is that API key tied only to your Account? or is it saved in a way that service (scheduled refresh) can use it too? (manual refreshes run as you but schedules run as the service)
3-If its only on your account recreate the connection....When you set it up again (make sure its created as a proper connection so scheduled refresh knows where to grab the API key) / it also could be API limit.
4-I am out of solutions (without error code), so if you Still failing....Grab the request ID and activity ID then raise a ticket with Microsoft Support (They can dig into the backend logs and will be helpful more than us (without error code)).
Let me know if you have another questions :).
Hi NiallFabric
It sounds like the scheduled flows are failing intermittently after the update to use an API key, even though manual runs continue to succeed. This pattern often indicates an authentication or connectivity issue in the cloud runtime environment rather than a problem with the flow logic itself. When flows run manually, they execute under your user context, which may have cached credentials or a more stable session, but scheduled runs use the service context, which might expose timing, rate limiting, or token expiration issues with the API key. Other potential causes include timeouts if the API is slow to respond, or transient errors like data merge or payload size issues that only appear during scheduled runs due to larger data volumes or concurrent execution. The lack of detailed error information in the CSV is typical for service-level errors. The best next steps are to check the flow’s run history in detail in Power Automate, verify that the API key has no restrictions or expiration, ensure the connector is configured to handle retries and timeouts, and monitor the connected system for any throttling or rate-limit errors that coincide with the scheduled runs.
Hi NiallFabric ,
May I ask if you have resolved this issue? Please let us know if you have any further issues, we are happy to help.
Thank you.
Hi, partially resolved, there seems to be an underlying issue with the new API key method that is making the (smaller than ever before dataflow) eat up more bandwidth.
I've started a Fabric trial and converted the flows to Gen2 flows. They have been running more consitently, and the Gen1s are a little better at manual syncs.
Quite a spendy soloution for a flow that has been larger and previously worked reasonably well with just Power BI pro.
With the trial upgrades, I am getting better error reporting which points to the intermittent failures always happening on one databoard at the source.
<URL> (500): WorkBookMessageException: MashupValueException.
The source platform support can't seem to figure out why but have tried a few things (with no luck). The fact that the Gen2 runs reasonably consistently suggests the API key and parameters may just be making the Gen1 pretty flakey with that platform.
Might need to bite the bullet and add the overhead, for what is a change of login method to a more secure form. Longterm we would move to a different soloution for the cost (even if just F2 and a user account for one dashboard).
Hi NiallFabric ,
I hope the information provided is helpful.I wanted to check whether you were able to resolve the issue with the provided solutions.Please let us know if you need any further assistance.
Thank you.
Hi NiallFabric ,
The 500 MashupValueException is coming from the source system, not Fabric. Gen1 is less resilient, which is why your scheduled refresh fails, while Gen2 handles retries and errors better - explaining why it’s more stable.
you can try below:
Run the failing query alone in Power Query to confirm it’s the root cause.Reduce payload size (filter/trim fields, avoid expanding large nested objects early).Check if the API key method enforces stricter throttling or bandwidth caps.If reliability is critical, stick with Gen2 (even a small F2 capacity) or stage the data in an intermediate DB/API wrapper.
Finally,you can try to raise support ticket using Microsoft's official support ticket channel using below link:
How to create a Fabric and Power BI Support ticket - Power BI | Microsoft Learn
they can confirm if the 500s are source-side.Thank you.
Hi NiallFabric ,
I wanted to check if you had a chance to raise a support ticket and resolve this issue? If so, please consider sharing the solution in the forum,this will help other members find the answer more easily.
Thank you.
