I have an Azure data explorer cluster with a few databases in it. When looking at the 'Cluster boundedness' tab in Insights it states that both CPU and Cache utilization is low. However the Ingestion utilization is high.
One would assume that with a CPU usage average of ~3.5% there would be little variation in query times. When performing a query it also states that the CPU load is low:
I've tried looking for documentation or guides on how to reduce or optimize the ingestion utilization without the need for vertical or horizontal scaling but haven't found anything concrete.
We're running on Standard_E4d_v4 SKU and utilizing streaming ingestion. According to the metrics the utilization looks low:

So my question is if there are any concrete ways to reduce the ingestion utilization while keeping near real-time queries with streaming ingestion or get more detailed information on why this metric stays high and impacts query performance.

