CloudWatch alarms in AWS - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ service_behavior

intermediate

Understanding CloudWatch Alarm States

What is the state of a CloudWatch alarm immediately after it is created but before any metric data is received?

A INSUFFICIENT_DATA

B OK

C ALARM

D DISABLED

💡 Hint

Think about what happens when there is no data to evaluate the alarm.

When a CloudWatch alarm is first created, it has no data to evaluate. Therefore, its state is INSUFFICIENT_DATA until metric data arrives and the alarm can evaluate its condition.

❓ Configuration

intermediate

Alarm Configuration for CPU Utilization

You want to create a CloudWatch alarm that triggers when the average CPU utilization of an EC2 instance exceeds 70% for 5 consecutive minutes. Which configuration below correctly sets the evaluation period and threshold?

A EvaluationPeriods=1, Threshold=70, Statistic=Average, Period=300

B EvaluationPeriods=1, Threshold=70, Statistic=Maximum, Period=300

C EvaluationPeriods=5, Threshold=70, Statistic=Sum, Period=60

D EvaluationPeriods=5, Threshold=70, Statistic=Average, Period=60

💡 Hint

Remember that the period is in seconds and evaluation periods count how many periods must breach the threshold.

To check if CPU utilization exceeds 70% for 5 consecutive minutes, set Period to 300 seconds (5 minutes) and EvaluationPeriods to 1. Statistic should be Average to measure average CPU usage.

❓ security

advanced

IAM Permissions for CloudWatch Alarm Actions

Which IAM permission is required to allow a user to create and manage CloudWatch alarms that send notifications to an SNS topic?

A cloudwatch:CreateAlarm and sns:Subscribe

B cloudwatch:PutMetricAlarm and sns:Publish

C cloudwatch:PutAlarm and sns:SendMessage

D cloudwatch:CreateMetricAlarm and sns:Notify

💡 Hint

Check the exact IAM actions for CloudWatch alarms and SNS publishing.

To create or update alarms, the user needs cloudwatch:PutMetricAlarm permission. To send notifications to SNS, sns:Publish permission is required.

❓ Architecture

advanced

Designing a Multi-Region Alarm System

You want to monitor an application deployed in two AWS regions and trigger a single alarm if the error rate exceeds 5% in either region. Which architecture best achieves this?

A Create CloudWatch alarms in each region and configure each to send notifications to separate SNS topics without aggregation.

B Create a single CloudWatch alarm in one region monitoring metrics from both regions directly.

C Create separate CloudWatch alarms in each region and use an SNS topic subscribed by a Lambda function that aggregates and triggers a global alarm.

D Use CloudWatch cross-region dashboards to view both regions but create alarms only in the primary region.

💡 Hint

Consider how CloudWatch alarms work regionally and how to aggregate alerts.

CloudWatch alarms are regional and cannot monitor metrics from other regions directly. Creating alarms in each region and aggregating notifications via SNS and Lambda allows triggering a single global alarm.

✅ Best Practice

expert

Optimizing CloudWatch Alarm Costs and Accuracy

You have many CloudWatch alarms monitoring various metrics. To reduce costs without losing critical alert accuracy, which approach is best?

A Increase the alarm evaluation period and number of evaluation periods to reduce alarm state changes and API calls.

B Set all alarms to use the Maximum statistic and a short period to catch spikes quickly.

C Delete alarms for low-priority metrics and rely on manual monitoring for those.

D Use composite alarms to combine multiple alarms and reduce the number of notifications.

💡 Hint

Think about combining alarms to reduce noise and cost.

Composite alarms combine multiple alarms into one, reducing the number of notifications and API calls, which lowers costs while maintaining alert accuracy.