Recent operator-sdk upgrade adjusted operator deployment limits. Those limits are enough for common usage, but in case of profiling the operator pod hits the memory limit, causing pod termination and respin. The memory limit needs to be raised to avoid restarts.