Exam AWS DevOps Engineer Professional topic 1 question 23 discussion

A: https://docs.aws.amazon.com/elasticloadbalancing/latest/classic/elb-cloudwatch-metrics.html https://docs.aws.amazon.com/codedeploy/latest/userguide/deployments-stop.html

A - correct B - this might work but has additonal overhead of a lambda and will depende on how frequently lambda is run. Although minimal but additonal cost of lambda. This won't give QUICKEST detection time. C - MinimumHealthyHosts may not be directly correlated with latency. Latency might be more due to network or other issues even if 100% of hosts are healthy. D - why do this when a ready made option is available https://docs.aws.amazon.com/elasticloadbalancing/latest/classic/elb-cloudwatch-metrics.html

A is the right answer , since it is readily available , we don't have to spend time creating the metrics from logs

Vote for A

We need QUICKEST detection time, not easiest deployment. Option A can only do per minute. Option D can go for second-level detection.

Which of the following methods allow for the QUICKEST detection time? Amazon CloudWatch Metrics: By default, CloudWatch metrics have a minimum resolution of 1 minute. This means that data points for metrics are collected at one-minute intervals. However, you can enable high-resolution custom metrics with a resolution of 1 second. Amazon CloudWatch Logs: CloudWatch Logs does not have a predefined minimum interval for log data collection. Log events can be sent to CloudWatch Logs in near real-time, allowing for timely analysis and monitoring.

This method allows for the quickest detection time since it relies on CloudWatch metrics provided by Elastic Load Balancing. These metrics are already being collected and aggregated by AWS, so there is no need to set up additional monitoring infrastructure. By creating a CloudWatch alarm on the average latency metric and setting a threshold, the DevOps Engineer can quickly detect when latency increases beyond a defined threshold and stop deployment until the service has recovered.

A- is Correct, it is quick as metrics could be updated every 1 minutes C- cannot be the answer https://docs.aws.amazon.com/codedeploy/latest/APIReference/API_MinimumHealthyHosts.html when we are talking about delay the loadbalancer could consider that endpoint as healthy and deployment will go through!

A. Use Amazon CloudWatch metrics provided by Elastic Load Balancing to calculate average latency. Alarm and stop deployment when latency increases beyond the defined threshold is the method that allows for the quickest detection time. When using Amazon CloudWatch metrics provided by Elastic Load Balancing, metrics are collected every minute, which is the most frequent metric collection interval available. This means that the latency metric will be updated frequently, allowing for a quick detection time when latency increases beyond the defined threshold. In contrast, the other options may have longer detection times as they involve additional steps such as collecting logs or waiting for a certain number of healthy hosts before rolling back a deployment.

the answer is A

A is the most straightforward solution. B: Lambda and Elastic Load Balancing, is manual solution. C and D look weird

The correct answer is A. When setting up CodeDeploy as a Deployment Provider you can specify the alarm based on a CloudWatch Metric (custom or built-in) and stop the deployment if an alarm threshold is exceeded.

A IS THE CORRECT ONE

I'll go with A

A is the way to go here

A, existing metric would be faster than going over the fuss of creating a new one

I also vote for A.