CompTIA Cloud+ (CV0-004)TroubleshootingHard
A cloud administrator is investigating an issue where users are reporting slow response times and occasional timeouts when accessing a web application hosted on multiple virtual machines behind a load balancer. Monitoring shows that CPU utilization on the application VMs is consistently low (under 30%), but network I/O metrics indicate high packet retransmission rates. What is the MOST likely cause of the performance degradation?
- AMisconfigured load balancer health checks.
- BNetwork bandwidth saturation on the VM network interfaces.
- CDatabase connection limits being reached by the application.
- DInsufficient memory allocated to the application VMs.
Show answer & explanationAnswer & explanation
Correct answer: B. Network bandwidth saturation on the VM network interfaces.
Low CPU but high packet retransmissions point to a network bottleneck, not a compute or application issue. High retransmission rates indicate that packets are being dropped or delayed significantly, often due to the network interface or underlying network infrastructure being unable to handle the traffic volume.
Why the other options are wrong
- A. Misconfigured health checks would lead to instances being unhealthy or removed, causing 5xx errors or connection refused, not high packet retransmission rates on *healthy* instances.
- C. Database connection limits would typically manifest as application errors (e.g., 'too many connections'), not directly as high packet retransmission rates at the network layer on the web application VMs.
- D. Insufficient memory typically leads to high swap usage, application crashes, or very high CPU due to garbage collection, not directly high packet retransmission with low CPU.
Network Bandwidth Saturation
A condition where the data transfer capacity of a network link or interface is fully utilized, leading to increased latency, packet loss, and retransmissions.
- Symptoms include high packet retransmission rates, increased latency, and timeouts.
- Often occurs when CPU/memory are normal, indicating a network-specific bottleneck.
- Can be mitigated by increasing network interface capacity or distributing traffic.
Memory trick: Low CPU, high retransmissions? Your network pipe is too small!