A SysOps Administrator is configuring a monitoring and remediation setup for a batch processing application running on Amazon EC2. The application publishes a custom metric named `QueueBacklog` to Amazon CloudWatch every 1 minute. The administrator needs to configure a CloudWatch alarm that meets the following requirements:
- The alarm must transition to the ALARM state if the queue backlog is greater than 1000 items for a continuous duration of 15 minutes.
- If there is a temporary network interruption preventing the application from publishing metrics, the alarm must default to the ALARM state to ensure visibility.
Which TWO of the following configurations are required to meet these requirements?
- Set the alarm Period to 1 minute and the Evaluation Periods to 15.Answer
- Configure the alarm's missing data treatment to treat missing data as breaching.Answer
- CEnable detailed monitoring on the EC2 instances to allow the custom metric to be published at 1-minute intervals.
- DCreate an Amazon EventBridge rule that monitors the QueueBacklog metric directly to trigger an Amazon SNS topic.
- EIncrease the Auto Scaling group cooldown period to 15 minutes to pause metric collection during instance scale-out.
Answer
The correct configurations are setting the alarm period to 1 minute with 15 evaluation periods, and configuring the missing data treatment to treat missing data as breaching.
To monitor a continuous 15-minute window for a metric published every 1 minute, the alarm needs to evaluate 15 individual 1-minute periods (or datapoints). Additionally, to force the alarm into the ALARM state when network issues prevent data publication, the missing data treatment must be set to 'breaching'.
Step-by-Step Solution
Key Concept
CloudWatch Alarm Evaluation and Missing Data Configuration