What causes Fabric Pipeline Error 3204?

The Databricks notebook or job encountered a runtime error such as a null reference, data mismatch, or out-of-memory condition. The cluster entered an unexpected state during job execution (e.g., driver crash or node loss). The job timed out before completing due to large data volumes or inefficient code. No error message is surfaced because the Databricks job failed silently or the error was swallowed by the cluster

How do I fix Fabric Pipeline Error 3204?

Step 1: Open the Databricks workspace and navigate to the Jobs or Clusters run history to find the specific run that failed and review its detailed logs.. Step 2: Check the Spark driver logs and event logs in Databricks for stack traces, out-of-memory errors, or node failures that correspond to the failed run.. Step 3: If the notebook caused the failure, reproduce the error by running the notebook manually in Databricks with the same parameters to isolate the root cause.. Step 4: Review the cluster configuration for resource adequacy — increase worker nodes or memory if the job is failing due to resource exhaustion.. Step 5: Add robust error handling and logging within the Databricks notebook itself so that failures produce actionable error messages surfaced back to the Fabric pipeline.

High severityrefreshMicrosoft Fabric →

Power BI Refresh Error:
Fabric Pipeline Error 3204

What does this error mean?

The Databricks job submitted by the Fabric pipeline failed during execution. Error messages can vary widely and may indicate an unexpected cluster state, notebook error, or other activity-level failure — sometimes with no explicit error message returned.

Common causes

1The Databricks notebook or job encountered a runtime error such as a null reference, data mismatch, or out-of-memory condition
2The cluster entered an unexpected state during job execution (e.g., driver crash or node loss)
3The job timed out before completing due to large data volumes or inefficient code
4No error message is surfaced because the Databricks job failed silently or the error was swallowed by the cluster

How to fix it

1Step 1: Open the Databricks workspace and navigate to the Jobs or Clusters run history to find the specific run that failed and review its detailed logs.
2Step 2: Check the Spark driver logs and event logs in Databricks for stack traces, out-of-memory errors, or node failures that correspond to the failed run.
3Step 3: If the notebook caused the failure, reproduce the error by running the notebook manually in Databricks with the same parameters to isolate the root cause.
4Step 4: Review the cluster configuration for resource adequacy — increase worker nodes or memory if the job is failing due to resource exhaustion.
5Step 5: Add robust error handling and logging within the Databricks notebook itself so that failures produce actionable error messages surfaced back to the Fabric pipeline.

Frequently asked questions

Why is there sometimes no error message with Error 3204?

Error 3204 is a generic catch-all for Databricks job execution failures. When the cluster itself crashes or the error is not properly propagated by Databricks back to Fabric, no detailed message is returned. In these cases, the Databricks run logs are the only source of truth.

How can I get better error details from this failure?

Add try-except blocks in your Databricks notebooks that log errors to a Delta table or raise exceptions with descriptive messages. This ensures that meaningful error information is captured even when the default pipeline error message is empty.