How Do You Alarm on EBS Disk Utilization Above 80% in CloudWatch?
Short answer
ACE Attach an IAM role with CloudWatchAgentServerPolicy, install and start the CloudWatch agent on the instances, then alarm on disk_used_percent greater than 80%.
Question
Which combination of steps must the CloudOps engineer take to meet these requirements? (Choose three.)
Key Takeaways
Tests whether you know that guest-OS disk metrics come only from the CloudWatch agent and are named disk_used_percent — the trap is assuming EBS or EC2 publishes a ready-made disk utilization metric or guessing metric names like disk_used and disk_free.
Amazon EBS volumes do not natively publish disk-space metrics, so monitoring disk utilization on EC2 Linux instances requires the CloudWatch agent plus an IAM role carrying CloudWatchAgentServerPolicy. The page confirms the three-step combination — role with CloudWatchAgentServerPolicy, agent installation, and a disk_used_percent alarm above 80% — as the correct answer (A, C, E).
Common mistake: Many candidates skip the IAM prerequisite or pick option D, which redundantly re-states the role requirement, or they choose option F and invent the metric names disk_used and disk_free instead of the real agent metric disk_used_percent.
Expert Analysis
Why the Answer Is Correct
EBS volumes expose volume-level metrics such as VolumeReadBytes and BurstBalance, but never filesystem occupancy, so the only way to see how full a Linux filesystem is on an EC2 instance is the CloudWatch agent. Option A supplies the required permission path: an IAM role carrying the AWS managed policy CloudWatchAgentServerPolicy attached to the instances, which lets the agent call cloudwatch:PutMetricData. Option C covers deployment, installing and starting the agent through AWS Systems Manager or the command line, which is the supported way to get the agent running on the fleet. Option E names the exact metric the agent publishes for this purpose, disk_used_percent, and compares it against the 80% threshold the company requires. Together these three steps produce a working alarm that fires when disk utilization exceeds 80%.
Why the Other Options Are Wrong
Option B attaches CloudWatchApplicationInsightsReadOnlyAccess, a read-only Application Insights policy that grants nothing the agent needs to publish metrics, so the agent would fail to send data. Option D overlaps with option A and is awkwardly framed as installing the agent "by using an IAM role"; it cannot coexist as an additional step once A is selected, and it does not describe installation at all. Option F cites disk_used and disk_free, metric names the CloudWatch agent does not emit for Linux filesystems, so an alarm built on them would sit in INSUFFICIENT_DATA indefinitely. Only disk_used_percent (with dimensions such as path, device, and fstype) reflects the percentage of space consumed.
Community Comment Notes
Bad_Mat summarized the reasoning directly, stating that "The correct combination of steps is A, C, and E." and spelling out the role with CloudWatchAgentServerPolicy, the agent installation step, and the disk_used_percent alarm at greater than 80%. That matches the official metric reference and the IAM requirements for the CloudWatch agent, and it is consistent with how the exam frames agent-based guest-OS monitoring. The comment reinforces that the IAM role is a separate, mandatory step from installation, which is exactly where most distractors try to trip candidates.
Official Reference
- https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/install-CloudWatch-Agent-on-EC2-Instance.html
- https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/create-iam-roles-for-cloudwatch-agent.html
- https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/metrics-collected-by-CloudWatch-agent.html
These links open the official vendor documentation in a new tab.
Exam Strategy
Whenever a question asks about memory, disk space, or swap on EC2, immediately think CloudWatch agent plus an IAM role with CloudWatchAgentServerPolicy — EC2 and EBS alone never provide those guest-OS metrics. Then match the exact metric name (disk_used_percent) rather than a plausible-sounding invention, and treat duplicate-looking options as a sign that only one of them belongs in the three-step answer.
Frequently Asked Questions
Why can't I just alarm on an EBS metric for disk utilization?
Why is option D wrong when it also mentions CloudWatchAgentServerPolicy?
Which exact metric does the CloudWatch agent report for disk usage?
Discussion 1 contributions
Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.
From public community threads 1
-
Bad_Mat 👍 1The correct combination of steps is A, C, and E.“A. Create an IAM role that includes the CloudWatchAgentServerPolicy AWS managed policy. Attach the role to the instances. C. Install and start the CloudWatch agent by using AWS Systems Manager or the command line. E. Configure a CloudWatch alarm to enter ALARM state when the disk_used_percent CloudWatch metric is greater than 80%.”, as noted in Exam AWS Certified SysOps Administrator - Associate topic 1.
Reader comments & corrections
No reader corrections yet — spotted an error or have a note? Share it below.
Related Analysis
Practice All SOA-C03 Questions
Access 100 questions with complete answers and detailed explanations.
View Full SOA-C03 Practice Test →