KodeKloud Notes

In this article, you’ll learn how to create an AWS Fault Injection Simulator (FIS) experiment template that simulates an Availability Zone (AZ) power interruption for an Amazon Aurora database. You will:

Identify the AZ of your Aurora writer instance
Select and configure the appropriate FIS scenario
Limit the blast radius with resource tags
Define advanced parameters for timing
Set up CloudWatch Logs for monitoring

1. Identify the Target AZ for Your Aurora Writer

Open the Amazon RDS console.
Select your Aurora cluster.
Note the Availability Zone (AZ) where the writer instance is running (e.g., us-east-1a).

The image shows an Amazon RDS dashboard displaying a list of databases with their identifiers, status, roles, engine type, region, and size. There is a notification about Aurora I/O-Optimized and a suggestion to consider Blue/Green Deployment.

2. Select the AZ Availability: Power Interruption Scenario

Navigate to the AWS Fault Injection Simulator (FIS) console.
Click Scenario Library to browse pre-built scenarios.
Choose AZ Availability: Power Interruption.

The image shows the AWS Fault Injection Simulator (FIS) interface, specifically the Scenario Library, with options for different resilience testing scenarios like "AZ Availability: Power Interruption" and "EC2 Stress: CPU."

3. Review Scenario Actions and Targets

The AZ Availability: Power Interruption template orchestrates the following actions in sequence:

Pause EC2 instance launches
Pause EBS I/O
Adjust Auto Scaling Group (ASG) scaling

Targets include IAM roles, EC2 instances, and EBS volumes in the specified AZ.

The image shows an AWS Fault Injection Simulator interface with actions and targets for a power interruption scenario, including "Pause-Instance-Launches" and "Pause-EBS-IO" actions. It highlights IAM roles and EBS volumes as targets, with a note about missing resource IDs.

4. Configure Scenario Parameters

Affected Availability Zone & IAM Role

Set Affected Availability Zone to your Aurora writer’s AZ (e.g., us-east-1a).
Specify the IAM role that FIS will assume to manage EC2 and EBS resources.

The image shows a configuration screen for selecting scenario parameters in a cloud service, specifically for setting parameters related to availability zone impairments and power interruptions.

Targeting Tags

Limit the blast radius by targeting only resources with predefined tags.

Tag Key	Tag Value
AZ-impairment	power-ready

Note

Ensure your EC2 instances and EBS volumes are already tagged with these keys and values before running the experiment.

Advanced Timing Parameters

Align these durations with your application’s RPO (Recovery Point Objective) and RTO (Recovery Time Objective).

Phase	Description	Duration
DNS Impact	Time for DNS propagation effects	2 minutes
Outage	Simulated power-off window	10 minutes
Recovery	Time to restore power	5 minutes

The image shows a configuration screen for setting parameters related to power impairment scenarios, including DNS impact, outage, and recovery durations.

5. Configure Logging to CloudWatch

Under Log group, search for fis.
Select the CloudWatch Logs group where FIS will send experiment logs.

6. Create and Review the Experiment Template

Verify all parameters and targets are correct.
Capture baseline (steady-state) metrics for comparison.
Click Create experiment template.

Warning

Chaos experiments can disrupt production workloads. Always test in a staging environment and monitor your application during the injection.

Links and References

Watch Video

Watch video content