Shell script: Find the bad burstable instances in AWS cloud

Following is the shell script to find bad burstable instances in AWS cloud architecture where all the credits are exhausted:

#!/bin/bash

# Get the list of all running instances
instances=$(aws ec2 describe-instances — query “Reservations[*].Instances[*].InstanceId” — output text)

# Filter the list to only include burstable instances
burstable_instances=$(echo “$instances” | grep -E “t2|a1”)

# For each burstable instance, check its CPU credits
for instance in $burstable_instances; do
cpu_credits=$(aws ec2 describe-instances-status — instance-ids $instance — query “InstanceStatuses[*].InstanceStatus.CpuCredits” — output text)

# If the CPU credits are below 10, then the instance is bad
if [ $cpu_credits -lt 10 ]; then
# Provision a new instance of the same type
new_instance=$(aws ec2 run-instances — image-id $(aws ec2 describe-instances — instance-ids $instance — query “Reservations[*].Instances[*].ImageId” — output text) — instance-type $(aws ec2 describe-instances — instance-ids $instance — query “Reservations[*].Instances[*].InstanceType” — output text) | grep InstanceId | awk ‘{print $2}’)

# Wait for the new instance to be running
while [ $(aws ec2 describe-instances — instance-ids $new_instance — query “Reservations[*].Instances[*].InstanceState.Name” — output text) != “running” ]; do
sleep 10
done

# Terminate the bad instance
aws ec2 terminate-instances — instance-ids $instance
fi
done

Above script gives a base idea about figuring out the bad instances using shell script. It needs to be further modified to cater specific scenario and requirements.