What causes Low memory situations?

High-demand periods with many large semantic models loaded or refreshing simultaneously consuming available node memory. A single very large semantic model consuming most of the available memory on a physical node, leaving insufficient headroom for concurrent operations. Capacity SKU undersized relative to the total memory footprint of all active semantic models. Node-level limit exceeded due to uneven distribution of large semantic models across physical Premium nodes

How do I fix Low memory situations?

Step 1: Retry the failed refresh operation — memory pressure is often transient, and the system's automatic load balancing typically resolves it quickly.. Step 2: If the error persists, reschedule the refresh to an off-peak time when fewer semantic models are active or being refreshed concurrently.. Step 3: Review the Premium Capacity Metrics app to identify memory utilization trends and determine if the capacity is consistently near its memory ceiling.. Step 4: Reduce the memory footprint of large semantic models by removing unused columns, aggregating data further, or splitting the model into smaller incremental refresh segments.. Step 5: If memory exhaustion is frequent, consider upgrading the Premium SKU or migrating memory-intensive models to a higher-capacity node or separate capacity.

Medium severitycapacity

Power BI Refresh Error:
Low memory situations

What does this error mean?

The Power BI Premium capacity temporarily ran out of memory during a high-demand period, causing refresh operations to fail. The system typically recovers automatically as load balancing redistributes semantic models across nodes.

Common causes

1High-demand periods with many large semantic models loaded or refreshing simultaneously consuming available node memory
2A single very large semantic model consuming most of the available memory on a physical node, leaving insufficient headroom for concurrent operations
3Capacity SKU undersized relative to the total memory footprint of all active semantic models
4Node-level limit exceeded due to uneven distribution of large semantic models across physical Premium nodes

How to fix it

1Step 1: Retry the failed refresh operation — memory pressure is often transient, and the system's automatic load balancing typically resolves it quickly.
2Step 2: If the error persists, reschedule the refresh to an off-peak time when fewer semantic models are active or being refreshed concurrently.
3Step 3: Review the Premium Capacity Metrics app to identify memory utilization trends and determine if the capacity is consistently near its memory ceiling.
4Step 4: Reduce the memory footprint of large semantic models by removing unused columns, aggregating data further, or splitting the model into smaller incremental refresh segments.
5Step 5: If memory exhaustion is frequent, consider upgrading the Premium SKU or migrating memory-intensive models to a higher-capacity node or separate capacity.

Frequently asked questions

What is the difference between a capacity-level memory error and a node-level limit exceeded error?

A capacity-level memory error means the overall Premium capacity is low on memory across all nodes. A 'node level limit exceeded' error is more specific — it means the physical node where a particular semantic model resides has hit its memory ceiling, which may not affect other models on different nodes.

Does Power BI automatically move models to free up memory?

Yes, Power BI Premium uses automatic load balancing and will evict inactive models from memory to free up space for active operations. However, if all models are active simultaneously, eviction may not be sufficient and refreshes can still fail.