What causes CosmosDbSqlApiPartitionKeyExceedStorage?

All documents being written have the same partition key value, causing a single logical partition to grow beyond 20 GB. The partition key chosen for the Cosmos DB container has low cardinality, concentrating data in a few partitions. A data quality issue caused all records to use a null or empty partition key value

How do I fix CosmosDbSqlApiPartitionKeyExceedStorage?

Review the partition key strategy for the Cosmos DB container — choose a high-cardinality partition key that distributes data evenly.. If the existing partition key causes hot partitions, create a new container with a better partition key and migrate data to it.. Inspect the source data for null or empty partition key values and fix them before the copy.. Consider using a synthetic partition key (e.g., combining two fields) if no single field has sufficient cardinality.

High severitydata source

Power BI Refresh Error:
CosmosDbSqlApiPartitionKeyExceedStorage

What does this error mean?

An ADF write to Cosmos DB failed because the target logical partition exceeded the 20 GB storage limit — each logical partition has a 20 GB maximum.

Common causes

1All documents being written have the same partition key value, causing a single logical partition to grow beyond 20 GB
2The partition key chosen for the Cosmos DB container has low cardinality, concentrating data in a few partitions
3A data quality issue caused all records to use a null or empty partition key value

How to fix it

1Review the partition key strategy for the Cosmos DB container — choose a high-cardinality partition key that distributes data evenly.
2If the existing partition key causes hot partitions, create a new container with a better partition key and migrate data to it.
3Inspect the source data for null or empty partition key values and fix them before the copy.
4Consider using a synthetic partition key (e.g., combining two fields) if no single field has sufficient cardinality.

Frequently asked questions

Does this error affect all pipeline runs or just the current one?

Depends on the root cause. A persistent misconfiguration fails every run; a transient issue may resolve on retry. Check the run history.

Can this error appear in Azure Data Factory and Microsoft Fabric pipelines?

Yes — the same connector errors appear in both ADF and Fabric Data Factory pipelines.

How do I see the full error detail for an ADF pipeline failure?

In ADF Monitor, click the failed run, then the failed activity. The detail pane shows the error code, message, and sub-error codes.

Will downstream Power BI datasets be affected when an ADF pipeline fails?

Yes — a dataset refreshing after the pipeline will use stale data or fail if the target table was cleared. The Power BI refresh may succeed while serving wrong data.

Official documentation: https://learn.microsoft.com/en-us/azure/data-factory/connector-troubleshoot-azure-cosmos-db