Hi all, We're using Azure Cosmos DB mirrored to Microsoft Fabric, exposing data through the SQL Analytics Endpoint, and connecting GoodData via the Azure Synapse Connector (I know this approach is not supported by the GD). We have several metrics working successfully against the same dataset, but some metrics fail with some generic error code:

{
  "status": 400,
  "detail": "An error has occurred while calculating the result",
  "reason": "java.lang.RuntimeException: Error occurred while consuming data."
}

also I'm sendig you the traceId from these errors:

eb86863d970303bf6662fd690bce3bb3

d568db0a6d6ba409bf4aea5717345c0d

e99345338317059fabc5429a72ca1272

Could I ask for assistance with resolving these issues?

Hi! Thanks for the detailed context. Regarding the specific traceIds you've shared, the knowledge sources I have access to don't contain logs or details for those specific traceIds (

eb86863d970303bf6662fd690bce3bb3

,

d568db0a6d6ba409bf4aea5717345c0d

,

e99345338317059fabc5429a72ca1272

). Only GoodData's internal support team can look those up in their logs. What we do know from similar cases: The generic

"An error has occurred while calculating the result"

error almost always hides a more specific underlying cause that's only visible in internal logs. Past cases have revealed root causes such as: • SQL execution failures (invalid column names, missing columns, bad identifiers) [Community Slack] • Insufficient database permissions [Community Slack] • Connection/throttling issues [Community Slack] • LDM mapping issues (e.g. missing primary keys, broken label references) Given your specific setup (Cosmos DB → Fabric Mirroring → SQL Analytics Endpoint → Synapse connector), there's an important known constraint: the Fabric SQL Analytics Endpoint only supports Microsoft Entra ID (Azure AD) authentication, while GoodData's native Synapse SQL connector uses SQL auth. This mismatch could be contributing to intermittent or partial failures — especially if some metrics succeed while others fail, depending on how queries are routed or cached. [Community Slack] Recommended next steps: 1. Share your traceIds with GoodData Support — provide the three traceIds along with your workspace ID and the approximate time of the failures. Support can decode the exact SQL/runtime error behind

"java.lang.RuntimeException: Error occurred while consuming data."

[API errors] 2. Check browser DevTools — open the Network tab, reproduce the failing metric, and look for any additional error detail in the

/executionResults

response. 3. Consider FlexConnect as a longer-term alternative — since your setup is officially unsupported, FlexConnect (GoodData's custom connector framework using Python + Apache Arrow Flight) would allow you to connect to the Fabric endpoint with Entra ID auth, which the native Synapse connector cannot do. [Community Slack] You can reach GoodData Support via the Community Slack or by opening a ticket at **support.gooddata.com**. 👍 0 button 👎 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups. 👀 Our Support team monitors all messages and will jump in if needed.

Hi Aleskander, Thanks for reaching out and sharing details on the issue. We are looking into this. I will get back to you as soon as possible.

Helo Aleksander, We traced the error and found the root cause in the backend logs:

com.microsoft.sqlserver.jdbc.SQLServerException: An error occurred during the current command
(Done status 0). JSON text is not properly formatted. Unexpected character '.' is found at position 0.

This is thrown by SQL Server itself. It probably means one of the columns being read (likely a JSON-typed field from the Cosmos DB → Fabric mirroring) contains a value that isn't valid JSON. That's why only some metrics fail: they're the ones touching that field/row. We'd recommend checking the affected column directly on the SQL Analytics Endpoint for malformed JSON values.

sorry for late replay. As you said issue was caused by our view which faultily handled the json values. It's working now. Thanks for the assistance ❤️