After deploying multiple copies of database servers, data scrambling is started on them to anonymize user data. A few minutes later, the systems administrator receives multiple complaints regarding…

The storage array is a reasonable concern for I/O-heavy workloads, but it should be investigated after the database driver layer - which sits between the application and storage - has been ruled out as the source of excessive requests.

Data scrambling workloads use database drivers to read, transform, and write large volumes of records, and misconfigured or inefficient drivers can produce excessive I/O requests, connection pool exhaustion, or aggressive locking that saturates shared infrastructure. Since CPU and memory are ruled out, the software layer between the application and storage - the database drivers - is the next logical bottleneck to inspect. Driver-level inefficiencies can create resource contention that cascades to other VMs on the same host or storage fabric.

A hardware load balancer distributes external network traffic and is not involved in the internal communication between data scrambling processes and the database servers.

Internet connection speed affects external-facing traffic only and is unrelated to the performance of internal VMs communicating with database servers on the local network.