A SysOps administrator is managing a fleet of EC2 instances running a custom application. They need to ensure that the application process is always running. If the process stops, they want to automatically restart it and receive a notification. Which AWS services should be combined to achieve this?
- AAWS Systems Manager State Manager to ensure the process is running and send notifications via SNS.
- BAmazon CloudWatch Alarms monitoring a custom metric, triggering an AWS Systems Manager Automation document to restart the process and then send an SNS notification.
- CAmazon CloudWatch Alarms monitoring a custom metric, triggering an SNS topic and an AWS Lambda function to restart the process.
- DAWS Config rules to detect if the process is stopped, triggering an AWS Lambda function to restart it.
Show answer & explanationAnswer & explanation
Correct answer: B. Amazon CloudWatch Alarms monitoring a custom metric, triggering an AWS Systems Manager Automation document to restart the process and then send an SNS notification.
The CloudWatch agent can send a custom metric indicating the application process status (e.g., 0 for stopped, 1 for running). A CloudWatch Alarm can then be configured to trigger when this metric indicates the process is stopped. This alarm can trigger an SSM Automation document, which can execute commands (like restarting the process) on the EC2 instance. The Automation document can also be configured to send an SNS notification upon completion or failure, providing both automated remediation and notification.
Why the other options are wrong
- A. Systems Manager State Manager can ensure a desired state (process running), but it's typically for periodic enforcement, not immediate, event-driven restarts from a metric alarm. It also doesn't directly integrate with CloudWatch Alarms for proactive detection.
- C. Lambda can restart processes if it has the necessary permissions and can remotely execute commands on the EC2 instance (e.g., via SSM Run Command integration), but an SSM Automation document is a more direct and robust way to manage instance-level processes from an alarm.
- D. AWS Config monitors resource *configurations*, not live application process status. It's not suitable for detecting if a process is running/stopped on an EC2 instance.
EC2 Process Auto-Remediation
Automate the monitoring and remediation of critical application processes on EC2 instances by combining custom CloudWatch metrics, alarms, and Systems Manager Automation documents.
- CloudWatch Agent collects custom process metrics.
- CloudWatch Alarms detect process failures.
- SSM Automation documents restart processes.
- SNS provides notifications.
Memory trick: Metric watches, Alarm screams, Automation fixes, SNS tells all.