AWS Certified Advanced Networking – Specialty (ANS-C01)Network Management and OperationsMedium
A global media company uses AWS for its streaming platform, which experiences significant and unpredictable spikes in traffic during major live events. Their current Auto Scaling Group for web servers is configured to scale based on CPU utilization. However, customers are reporting buffering and slow load times during these spikes, even before CPU utilization reaches the scaling threshold. The network team suspects that the instances are being overwhelmed by network traffic before the CPU-based scaling can react. Which CloudWatch metric should be used as the primary scaling trigger to ensure the Auto Scaling Group reacts more quickly and effectively to network-intensive traffic spikes?
- AEC2/DiskReadBytes (Bytes).
- BEC2/CPUUtilization (Percent).
- CEC2/NetworkPacketOut (Count).
- DEC2/NetworkIn (Bytes).
Show answer & explanationAnswer & explanation
Correct answer: D. EC2/NetworkIn (Bytes).
The 'NetworkIn' metric (specifically 'NetworkIn' for incoming bytes) directly measures the amount of incoming network traffic to an EC2 instance. For a streaming platform experiencing buffering and slow load times due to traffic spikes, scaling based on incoming network bytes would allow the Auto Scaling Group to react much faster to network-intensive loads, provisioning new instances before the existing ones become saturated.
Why the other options are wrong
- A. DiskReadBytes relates to disk I/O, which is less likely to be the primary bottleneck for incoming streaming traffic compared to network throughput.
- B. CPUUtilization is currently used but is reacting too slowly, indicating it's not the primary bottleneck during network-intensive spikes.
- C. NetworkPacketOut measures outgoing packets, but the issue described (buffering, slow load times) points to instances being overwhelmed by *incoming* traffic before processing and sending it out.
Network-Based Auto Scaling
Using Amazon CloudWatch metrics like `NetworkIn` to trigger Auto Scaling Group adjustments, allowing applications to scale based on network traffic load rather than just CPU utilization.
- Reacts to incoming network traffic.
- Crucial for streaming, gaming, or high-ingest applications.
- Prevents network saturation before CPU becomes a bottleneck.
Memory trick: When the network's flooding, 'NetworkIn' is your scaling warning.