Power BI dataset refresh failures

A Power BI dataset built on Dynamics 365 data (Dataverse, Business Central, or Finance and Operations) refreshes on a schedule or on demand, pulling fresh data through either a cloud connection or an on-premises data gateway. Refresh failures are usually about the connection or the credential, not the report design itself.

"Gateway timeout" / "The operation has timed out"

Symptom. A scheduled refresh fails after running for a long time, often the same duration every time.

Cause. The query against the source (a large Dataverse or Business Central table, a complex DAX-driven query) is taking longer than the gateway's or the refresh's timeout allows — Power BI Pro has a lower refresh-duration ceiling than Premium/Fabric capacity.

Fix. Reduce the data volume pulled per refresh (filter to a relevant date range, use incremental refresh instead of a full reload), or move the dataset to a capacity with a higher timeout ceiling if the underlying data volume genuinely requires it.

Prevention. Design incremental refresh policies from the start on any table large enough to be a timeout risk — refreshing only the changed partition of data is both faster and less likely to hit a timeout than reloading everything every time.

"Gateway not found" / "No gateway is configured for this data source"

Symptom. A refresh that depends on an on-premises or virtual-network-gated data source fails immediately, without attempting any query.

Cause. The on-premises data gateway the dataset depends on is offline, was uninstalled, or the data source's credentials inside the gateway configuration were never mapped after a dataset was published or moved to a new workspace.

Fix. Confirm the gateway is running and check the gateway's data source mapping in the Power BI admin portal — every data source a dataset needs must be explicitly mapped to a gateway data source, not just have the gateway installed somewhere.

Prevention. Run the gateway on a server with monitoring and automatic restart on failure, not a single individual's workstation — a gateway machine going to sleep or being rebooted for Windows updates is one of the most common causes of "it worked yesterday" refresh failures.

"Credentials could not be validated" / "You don't have access to this data source"

Symptom. A previously-working refresh starts failing citing a credential problem.

Cause. The account whose credentials are stored for the data source connection had its password changed, its MFA requirement changed, or its access to the underlying Dataverse/Business Central/F&O environment revoked or reduced.

Fix. Re-enter and re-validate the data source credentials in the dataset's settings, ideally using a dedicated service account rather than a named individual's credentials.

Prevention. Use a service account with a non-expiring or centrally-managed credential for scheduled refreshes, and exclude it from personal password-rotation and MFA policies that assume an interactive human is behind the sign-in.

"Something went wrong. Please try refreshing again later" (a generic failure)

Symptom. A refresh fails with a message that gives no specific cause.

Cause. This generic message covers a wide range of underlying issues — a transient service issue on Power BI's side, a query that changed shape unexpectedly (a column renamed or removed at the source), or a capacity/throttling limit reached.

Fix. Check the detailed refresh history (not just the summary status) in the dataset settings for the underlying error, which is usually more specific than the generic banner message. If the source schema changed, update the Power Query steps to match.

Prevention. Alert on refresh failures explicitly (a Power Automate flow watching the Power BI REST API, or a Premium capacity's built-in monitoring) rather than relying on someone noticing a stale report — since a failed refresh leaves the last-good data visible with no obvious visual sign anything is wrong.

Where to go next

Refresh strategy sits alongside the wider platform choice covered in Power BI incremental refresh, Power BI gateway for on-prem data, and Power BI deployment pipelines. For the newer alternative to a scheduled refresh entirely, see Microsoft Fabric and Dynamics 365 on DirectLake.