ANS-C01 Question 471
Single answerYour organization has deployed a multi-tier application on AWS, with an Application Load Balancer (ALB) in front of the web servers and an Amazon RDS instance for the database. The application team wants to be alerted if the ALB’s 5XX error rate exceeds 5% for a sustained period of 5 minutes. What is the most efficient way to achieve this?
- A
Create a CloudWatch alarm on the ALB’s HTTPCode_ELB_5XX_Count metric, set a threshold for 5% error rate, and configure the alarm with a 5-minute evaluation period.
- B
Create a CloudWatch alarm on the ALB’s HTTPCode_ELB_5XX_Count metric, calculate the error rate using a math expression, and configure the alarm with a 5-minute evaluation period.
- C
Create an AWS Lambda function to monitor the ALB logs in Amazon S3, calculate the 5XX error rate, and send an alert via Amazon SNS if the rate exceeds 5% for 5 minutes.
- D
Enable AWS WAF on the ALB and configure a rate-based rule to monitor and alert on 5XX error rates exceeding 5%.
Show answer and explanation
Correct answer: B
Explanation
The ALB’s HTTPCode_ELB_5XX_Count metric in CloudWatch provides the count of 5XX errors, but it does not directly provide the error rate as a percentage. To calculate the error rate, you need to use a math expression that divides the 5XX error count by the total RequestCount. Configuring a CloudWatch alarm with this math expression and setting it to trigger when the error rate exceeds 5% for a sustained period of 5 minutes is the most efficient and native solution for this requirement.
- A. Incorrect.
This option incorrectly uses the HTTPCode_ELB_5XX_Count metric directly to calculate a percentage. CloudWatch metrics are absolute counts, so you need to convert them to a percentage using a math expression.
- B. Correct.
This is the correct option. You can use a math expression in CloudWatch to calculate the 5XX error rate as a percentage (e.g., dividing HTTPCode_ELB_5XX_Count by RequestCount) and configure an alarm to trigger when the rate exceeds 5% for 5 minutes.
- C. Incorrect.
While technically feasible, this option is inefficient and overly complex. Monitoring logs with a Lambda function requires custom code, more maintenance, and higher latency compared to using CloudWatch alarms.
- D. Incorrect.
AWS WAF rate-based rules are designed to limit request rates, not to monitor error rates. This option does not meet the requirement of alerting based on the 5XX error rate.