High CPU after bulk insert


We have recently observed an issue with our 3 node cockroach cluster, where CPU usage of one of the node is very high after a bulk insert activity (around 40k records) in one of the table. This table has a sequence number as primary key. After the insert, following are the ranges and lease holders:

After the bulk insert, CPU usage of Node 1 is ~80% and the number of active distributed queries also high.

CPU profile of the node which has high CPU usage:

Appreciate any suggestions on what could cause the issue high CPU after adding more data?

It seems like all of the CPU utilization is servicing scan requests. I anticipate that this is correlated with query load. Is it possible that the queries being run are full table scans?

Looking at the EXPLAIN for the queries that you’re running might help to track this down.