Lesson 5 · Scaling & load balancing

Scaling and load balancing

Put two servers in two AZs behind an Application Load Balancer, let an Auto Scaling group keep them alive and scale on CPU, and watch it replace a server you kill.

40 min read · intermediate · hands-on

By the end of this lesson you'll be able to

0 of 6 complete

Describe a server once in a launch template
Run an Auto Scaling group with min, max and desired capacity across two AZs
Put an Application Load Balancer, a listener and a target group in front of it
Lock the servers down so only the load balancer can reach them
Watch the group replace an instance that dies, and scale on CPU with target tracking
Choose between ALB and NLB, and know where TLS certificates from ACM fit

From one server to a fleet

The server from lesson 3 had two problems: if it died, the site died, and if traffic tripled, it couldn't grow. This lesson builds the standard fix, in the two-AZ VPC from lesson 4, in a real account (all of it deleted afterwards):

internet → Application Load Balancer (public subnets, 2 AZs)
             → listener HTTP:80 → target group (health check GET /)
                → Auto Scaling group: 2–4 × t3.micro from a launch template, spread across AZs

Two security groups lock it down:

Group	Inbound
`academy-alb-sg`	port 80 from `0.0.0.0/0`
`academy-web-sg`	port 80 only from `academy-alb-sg`

$ aws ec2 authorize-security-group-ingress --group-id $WEBSG --protocol tcp --port 80 --source-group $ALBSG

aws ec2 authorize-security-group-ingress --group-id $WEBSG --protocol tcp --port 80 --source-group $ALBSG

awsthe AWS command line

A security group can use another security group as its source, so nobody can skip the load balancer and hit a server directly.

One server, described once

A launch template holds everything from lesson 3's wizard (AMI, instance type, security group, user data, IMDSv2) so a machine can start servers without you:

Launch template academy-web-lt, version 1 (default), instance details: AMI ami-04394a3a96907e9ec, instance type t3.micro, security group sg-0dcb73ef50b44a171, no key pair. — (1) The template's name. (2) Its instance type. No key pair: these servers are never logged into by hand.© Screenshot: DevOps Academy

The user data this time also asks the metadata service which AZ the server is in, so you can see the load balancer spreading requests:

#!/bin/bash
dnf install -y nginx
TOKEN=$(curl -s -X PUT http://169.254.169.254/latest/api/token -H "X-aws-ec2-metadata-token-ttl-seconds: 60")
AZ=$(curl -s -H "X-aws-ec2-metadata-token: $TOKEN" http://169.254.169.254/latest/meta-data/placement/availability-zone)
echo "<h1>Hello from $(hostname) in $AZ</h1>" > /usr/share/nginx/html/index.html
systemctl enable --now nginx

Templates are versioned: change the AMI, create version 2, and roll it out (an instance refresh replaces the servers a few at a time).

The Auto Scaling group

Auto Scaling group academy-asg capacity overview: desired capacity 2, scaling limits 2 to 4, status At desired capacity. Details tab: launch template academy-web-lt, version Latest, t3.micro. Network: Availability Zones ap-south-1a and ap-south-1b with two subnets, Availability Zone distribution balanced best effort. — (1) Desired capacity 2. (2) Scaling limits: never fewer than 2, never more than 4.© Screenshot: DevOps Academy

Setting	Meaning
Min	the floor: the group never goes below it, even if every server dies
Desired	how many it runs right now; scaling policies change this number
Max	the ceiling, which also caps your bill

Give the group subnets in two or more AZs and it spreads instances evenly, so losing a whole AZ leaves half the fleet running:

Instance management tab: two instances, both InService, t3.micro, launched from academy-web-lt, one in aps1-az3 (ap-south-1b) and one in aps1-az1 (ap-south-1a), both Healthy. — Two instances, two AZs, both InService and Healthy.© Screenshot: DevOps Academy

The load balancer

An Application Load Balancer (ALB) takes requests on a listener (here HTTP:80), checks its rules and forwards to a target group. The target group health-checks every server (GET / every 10 seconds) and only sends traffic to the healthy ones:

Load balancer academy-alb: type Application, internet-facing, Active, in two subnets in ap-south-1a and ap-south-1b, DNS name academy-alb-2056515812.ap-south-1.elb.amazonaws.com. Listeners and rules: HTTP:80. — (1) The DNS name clients use; an ALB has no fixed IP. (2) The HTTP:80 listener.© Screenshot: DevOps Academy

The load balancer's Resource map: Listeners (1) HTTP:80 with 1 rule; Rules: default priority, forward to target group; Target groups: academy-web-tg, instance HTTP, 2 targets, 2 healthy; Targets: i-00b8678ed097303f9 and i-02007cc1f591c06e4 on port 80, both Healthy. — (1) Listener → rule → (2) target group → (3) targets: the whole request path, and the health of every hop.© Screenshot: DevOps Academy

Target group academy-web-tg: HTTP port 80, load balancer academy-alb, 2 total targets, 2 Healthy, 0 Unhealthy. Registered targets: two academy-web instances on port 80 in ap-south-1a and ap-south-1b, both Healthy. — Both targets healthy. The Auto Scaling group registers new instances here by itself.© Screenshot: DevOps Academy

Requests take turns between the AZs:

$ for i in 1 2 3 4; do curl -s http://academy-alb-2056515812.ap-south-1.elb.amazonaws.com; done
<h1>Hello from ip-10-0-23-40.ap-south-1.compute.internal in ap-south-1b</h1>
<h1>Hello from ip-10-0-23-40.ap-south-1.compute.internal in ap-south-1b</h1>
<h1>Hello from ip-10-0-10-240.ap-south-1.compute.internal in ap-south-1a</h1>
<h1>Hello from ip-10-0-10-240.ap-south-1.compute.internal in ap-south-1a</h1>

for i in 1 2 3 4; do curl -s http://academy-alb-2056515812.ap-south-1.elb.amazonaws.com; done

forloop over a list
dostart of what the loop runs each time
curlmake an HTTP request
-ssilent: no progress bar
doneend of the loop

Kill a server

Terminate one instance by hand, as if its hardware failed:

$ aws ec2 terminate-instances --instance-ids i-0dca3be571b8075be

aws ec2 terminate-instances --instance-ids i-0dca3be571b8075be

awsthe AWS command line

Auto Scaling group Activity history: (1) Terminating EC2 instance i-0dca3be571b8075be, waiting for ELB connection draining, because an EC2 health check indicated it was terminated; (2) Launching a new EC2 instance i-02007cc1f591c06e4, in response to an unhealthy instance needing to be replaced, at the same second. — (1) The dead instance is taken out of service and drained. (2) A replacement is launched the same second, with no one paged.© Screenshot: DevOps Academy

The load balancer stops sending traffic to the dying server (connection draining lets in-flight requests finish), the group launches a replacement from the template, and once its health check passes, traffic returns to two servers. The curl output above already includes the new server, ip-10-0-10-240.

Scaling on load

Automatic scaling tab: one dynamic scaling policy, cpu-50, type target tracking, enabled, as required to maintain average CPU utilization at 50, add or remove capacity units as required, instances need 300 seconds to warm up. — A target tracking policy: keep average CPU near 50% by changing desired capacity between min and max.© Screenshot: DevOps Academy

Policy	How it works	Use for
Target tracking	keep a metric near a target, like a thermostat	most apps: CPU, requests per target
Step scaling	add or remove N instances at alarm thresholds	when you need precise steps
Scheduled	change capacity at set times	known patterns: office hours, a sale at 9 a.m.

Scaling takes minutes (boot, user data, health checks), so it handles growth, not a sudden spike. Keep a little headroom in min.

ALB or NLB

	Application Load Balancer	Network Load Balancer
Layer	7: HTTP and HTTPS	4: TCP, UDP, TLS
Routing	by path, host, header	by port
IP	DNS name only	static IPs per AZ
Use for	web apps and APIs	very high throughput, non-HTTP protocols, fixed IPs

For HTTPS, request a free public certificate from AWS Certificate Manager (ACM) for your domain, add an HTTPS:443 listener with it, and redirect HTTP:80 to HTTPS. ACM renews the certificate automatically.

Clean up

Listener rules choose the target

An ALB listener evaluates rules from the lowest priority number upward. A path condition can route /images/* to one target group and a host condition can route api.example.com to another. The default action is the fallback when neither condition matches.

$ aws elbv2 create-rule --listener-arn "$LISTENER" --priority 10 \
    --conditions Field=path-pattern,Values='/images/*' \
    --actions Type=forward,TargetGroupArn="$IMAGES_TG"
$ aws elbv2 describe-rules --listener-arn "$LISTENER" \
    --query 'Rules[].{Priority:Priority,Conditions:Conditions}' --output table

aws elbv2 create-rule --listener-arn "$LISTENER" --priority 10 \

awsthe AWS command line

Sticky sessions and safe refreshes

Stickiness keeps a browser on one target, which can help a legacy in-memory session but causes uneven load. Prefer an external session store and stateless targets. Instance refresh replaces an Auto Scaling group gradually after a launch-template change; lifecycle hooks pause launch or termination so work can drain, and warm pools keep pre-initialised instances ready for bursts.

$ aws elbv2 modify-target-group-attributes --target-group-arn "$TG" \
    --attributes Key=stickiness.enabled,Value=true Key=stickiness.lb_cookie.duration_seconds,Value=300
$ aws autoscaling start-instance-refresh --auto-scaling-group-name "$ASG" \
    --preferences MinHealthyPercentage=90,InstanceWarmup=120

aws elbv2 modify-target-group-attributes --target-group-arn "$TG" \

awsthe AWS command line

The load balancer is billed per hour even with no traffic, so delete everything, in order:

$ aws autoscaling delete-auto-scaling-group --auto-scaling-group-name academy-asg --force-delete
$ aws elbv2 delete-load-balancer --load-balancer-arn $ALB && aws elbv2 wait load-balancers-deleted --load-balancer-arns $ALB
$ aws elbv2 delete-target-group --target-group-arn $TG
$ aws ec2 delete-launch-template --launch-template-name academy-web-lt
$ aws ec2 delete-security-group --group-id $WEBSG && aws ec2 delete-security-group --group-id $ALBSG

aws autoscaling delete-auto-scaling-group --auto-scaling-group-name academy-asg --force-delete

awsthe AWS command line

aws elbv2 delete-load-balancer --load-balancer-arn $ALB && aws elbv2 wait load-balancers-deleted --load-balancer-arns $ALB

&&run the next command only if this one worked

--force-delete terminates the group's instances too. The web security group can only be deleted once those instances are gone.

Mini mission

A self-healing web tier

Behind an Application Load Balancer, run an Auto Scaling group of 2 t3.micro instances across two AZs, reachable only through the load balancer. Show requests answered from both AZs, kill one instance, watch it be replaced, then delete everything.

Input: An AWS account, the AWS CLI, and a VPC with two public subnets (the lesson 4 VPC, or the default VPC)Output: curl answers from two AZs, an Activity entry replacing the instance you killed, and no load balancer or group left

Reveal one safe solution

$ V=vpc-…; SUBS="subnet-… subnet-…"                       # two public subnets in different AZs
$ ALBSG=$(aws ec2 create-security-group --vpc-id $V --group-name academy-alb-sg --description "HTTP to the ALB" --query GroupId --output text)
$ aws ec2 authorize-security-group-ingress --group-id $ALBSG --protocol tcp --port 80 --cidr 0.0.0.0/0
$ WEBSG=$(aws ec2 create-security-group --vpc-id $V --group-name academy-web-sg --description "HTTP from the ALB only" --query GroupId --output text)
$ aws ec2 authorize-security-group-ingress --group-id $WEBSG --protocol tcp --port 80 --source-group $ALBSG
$ AMI=$(aws ssm get-parameter --name /aws/service/ami-amazon-linux-latest/al2023-ami-kernel-default-x86_64 --query Parameter.Value --output text)
$ aws ec2 create-launch-template --launch-template-name academy-web-lt --launch-template-data \
    "{\"ImageId\":\"$AMI\",\"InstanceType\":\"t3.micro\",\"SecurityGroupIds\":[\"$WEBSG\"],\"UserData\":\"$(base64 -w0 user-data.sh)\",\"MetadataOptions\":{\"HttpTokens\":\"required\"}}"
$ TG=$(aws elbv2 create-target-group --name academy-web-tg --protocol HTTP --port 80 --vpc-id $V --health-check-path / --query 'TargetGroups[0].TargetGroupArn' --output text)
$ ALB=$(aws elbv2 create-load-balancer --name academy-alb --type application --subnets $SUBS --security-groups $ALBSG --query 'LoadBalancers[0].LoadBalancerArn' --output text)
$ aws elbv2 create-listener --load-balancer-arn $ALB --protocol HTTP --port 80 --default-actions Type=forward,TargetGroupArn=$TG
$ aws autoscaling create-auto-scaling-group --auto-scaling-group-name academy-asg --launch-template LaunchTemplateName=academy-web-lt,Version='$Latest' \
    --min-size 2 --max-size 4 --desired-capacity 2 --vpc-zone-identifier "$(echo $SUBS | tr ' ' ,)" \
    --target-group-arns $TG --health-check-type ELB --health-check-grace-period 60
$ DNS=$(aws elbv2 describe-load-balancers --load-balancer-arns $ALB --query 'LoadBalancers[0].DNSName' --output text)
$ for i in 1 2 3 4; do curl -s http://$DNS; done            # after a few minutes: both AZs answer
$ aws ec2 terminate-instances --instance-ids <one of the group's instances>
$ aws autoscaling describe-scaling-activities --auto-scaling-group-name academy-asg --max-items 2 --query 'Activities[].Description'

aws ec2 authorize-security-group-ingress --group-id $ALBSG --protocol tcp --port 80 --cidr 0.0.0.0/0

awsthe AWS command line

for i in 1 2 3 4; do curl -s http://$DNS; done

forloop over a list
dostart of what the loop runs each time
curlmake an HTTP request
-ssilent: no progress bar
doneend of the loop

aws autoscaling describe-scaling-activities --auto-scaling-group-name academy-asg --max-items 2 --query 'Activities[].Description'

--querypick fields from the JSON response

user-data.sh is the script from this lesson; the subnets must give instances public IPs so they can download nginx (if not, turn it on: aws ec2 modify-subnet-attribute --subnet-id … --map-public-ip-on-launch). Clean up with the commands above. An ALB plus two t3.micro for half an hour costs a few cents.

Try it yourself

0 of 6 steps done

Run each task in your own account or terminal, then tick it off.

Create a launch template with the nginx user data from lesson 3.
Create an Auto Scaling group of 2 instances across two public subnets, with ELB health checks.
Put an Application Load Balancer in front and curl its DNS name a few times. Which hosts answer?
Terminate one instance by hand and watch the Activity tab.
Add a target tracking policy at 50% CPU and read what it will do.
Delete the group, load balancer, target group, launch template and security groups.

Knowledge check

Quick check

Question 1 of 5

What does an Auto Scaling group do when one of its instances fails its health check?

Go deeper

Official reading for this lesson

Or get every quiz answer right and it completes itself.

Lesson 5 · Scaling & load balancing

Scaling and load balancing

Put two servers in two AZs behind an Application Load Balancer, let an Auto Scaling group keep them alive and scale on CPU, and watch it replace a server you kill.

40 min read · intermediate · hands-on

By the end of this lesson you'll be able to

0 of 6 complete

Describe a server once in a launch template
Run an Auto Scaling group with min, max and desired capacity across two AZs
Put an Application Load Balancer, a listener and a target group in front of it
Lock the servers down so only the load balancer can reach them
Watch the group replace an instance that dies, and scale on CPU with target tracking
Choose between ALB and NLB, and know where TLS certificates from ACM fit

From one server to a fleet

internet → Application Load Balancer (public subnets, 2 AZs)
             → listener HTTP:80 → target group (health check GET /)
                → Auto Scaling group: 2–4 × t3.micro from a launch template, spread across AZs

Two security groups lock it down:

Group	Inbound
`academy-alb-sg`	port 80 from `0.0.0.0/0`
`academy-web-sg`	port 80 only from `academy-alb-sg`

$ aws ec2 authorize-security-group-ingress --group-id $WEBSG --protocol tcp --port 80 --source-group $ALBSG

aws ec2 authorize-security-group-ingress --group-id $WEBSG --protocol tcp --port 80 --source-group $ALBSG

awsthe AWS command line

A security group can use another security group as its source, so nobody can skip the load balancer and hit a server directly.

One server, described once

A launch template holds everything from lesson 3's wizard (AMI, instance type, security group, user data, IMDSv2) so a machine can start servers without you:

The user data this time also asks the metadata service which AZ the server is in, so you can see the load balancer spreading requests:

#!/bin/bash
dnf install -y nginx
TOKEN=$(curl -s -X PUT http://169.254.169.254/latest/api/token -H "X-aws-ec2-metadata-token-ttl-seconds: 60")
AZ=$(curl -s -H "X-aws-ec2-metadata-token: $TOKEN" http://169.254.169.254/latest/meta-data/placement/availability-zone)
echo "<h1>Hello from $(hostname) in $AZ</h1>" > /usr/share/nginx/html/index.html
systemctl enable --now nginx

Templates are versioned: change the AMI, create version 2, and roll it out (an instance refresh replaces the servers a few at a time).

The Auto Scaling group

Setting	Meaning
Min	the floor: the group never goes below it, even if every server dies
Desired	how many it runs right now; scaling policies change this number
Max	the ceiling, which also caps your bill

Give the group subnets in two or more AZs and it spreads instances evenly, so losing a whole AZ leaves half the fleet running:

The load balancer

Requests take turns between the AZs:

$ for i in 1 2 3 4; do curl -s http://academy-alb-2056515812.ap-south-1.elb.amazonaws.com; done
<h1>Hello from ip-10-0-23-40.ap-south-1.compute.internal in ap-south-1b</h1>
<h1>Hello from ip-10-0-23-40.ap-south-1.compute.internal in ap-south-1b</h1>
<h1>Hello from ip-10-0-10-240.ap-south-1.compute.internal in ap-south-1a</h1>
<h1>Hello from ip-10-0-10-240.ap-south-1.compute.internal in ap-south-1a</h1>

for i in 1 2 3 4; do curl -s http://academy-alb-2056515812.ap-south-1.elb.amazonaws.com; done

forloop over a list
dostart of what the loop runs each time
curlmake an HTTP request
-ssilent: no progress bar
doneend of the loop

Kill a server

Terminate one instance by hand, as if its hardware failed:

$ aws ec2 terminate-instances --instance-ids i-0dca3be571b8075be

aws ec2 terminate-instances --instance-ids i-0dca3be571b8075be

awsthe AWS command line

Scaling on load

Policy	How it works	Use for
Target tracking	keep a metric near a target, like a thermostat	most apps: CPU, requests per target
Step scaling	add or remove N instances at alarm thresholds	when you need precise steps
Scheduled	change capacity at set times	known patterns: office hours, a sale at 9 a.m.

Scaling takes minutes (boot, user data, health checks), so it handles growth, not a sudden spike. Keep a little headroom in min.

ALB or NLB

	Application Load Balancer	Network Load Balancer
Layer	7: HTTP and HTTPS	4: TCP, UDP, TLS
Routing	by path, host, header	by port
IP	DNS name only	static IPs per AZ
Use for	web apps and APIs	very high throughput, non-HTTP protocols, fixed IPs

Clean up

Listener rules choose the target

$ aws elbv2 create-rule --listener-arn "$LISTENER" --priority 10 \
    --conditions Field=path-pattern,Values='/images/*' \
    --actions Type=forward,TargetGroupArn="$IMAGES_TG"
$ aws elbv2 describe-rules --listener-arn "$LISTENER" \
    --query 'Rules[].{Priority:Priority,Conditions:Conditions}' --output table

aws elbv2 create-rule --listener-arn "$LISTENER" --priority 10 \

awsthe AWS command line

Sticky sessions and safe refreshes

$ aws elbv2 modify-target-group-attributes --target-group-arn "$TG" \
    --attributes Key=stickiness.enabled,Value=true Key=stickiness.lb_cookie.duration_seconds,Value=300
$ aws autoscaling start-instance-refresh --auto-scaling-group-name "$ASG" \
    --preferences MinHealthyPercentage=90,InstanceWarmup=120

aws elbv2 modify-target-group-attributes --target-group-arn "$TG" \

awsthe AWS command line

The load balancer is billed per hour even with no traffic, so delete everything, in order:

$ aws autoscaling delete-auto-scaling-group --auto-scaling-group-name academy-asg --force-delete
$ aws elbv2 delete-load-balancer --load-balancer-arn $ALB && aws elbv2 wait load-balancers-deleted --load-balancer-arns $ALB
$ aws elbv2 delete-target-group --target-group-arn $TG
$ aws ec2 delete-launch-template --launch-template-name academy-web-lt
$ aws ec2 delete-security-group --group-id $WEBSG && aws ec2 delete-security-group --group-id $ALBSG

aws autoscaling delete-auto-scaling-group --auto-scaling-group-name academy-asg --force-delete

awsthe AWS command line

aws elbv2 delete-load-balancer --load-balancer-arn $ALB && aws elbv2 wait load-balancers-deleted --load-balancer-arns $ALB

&&run the next command only if this one worked

--force-delete terminates the group's instances too. The web security group can only be deleted once those instances are gone.

Mini mission

A self-healing web tier

Reveal one safe solution

$ V=vpc-…; SUBS="subnet-… subnet-…"                       # two public subnets in different AZs
$ ALBSG=$(aws ec2 create-security-group --vpc-id $V --group-name academy-alb-sg --description "HTTP to the ALB" --query GroupId --output text)
$ aws ec2 authorize-security-group-ingress --group-id $ALBSG --protocol tcp --port 80 --cidr 0.0.0.0/0
$ WEBSG=$(aws ec2 create-security-group --vpc-id $V --group-name academy-web-sg --description "HTTP from the ALB only" --query GroupId --output text)
$ aws ec2 authorize-security-group-ingress --group-id $WEBSG --protocol tcp --port 80 --source-group $ALBSG
$ AMI=$(aws ssm get-parameter --name /aws/service/ami-amazon-linux-latest/al2023-ami-kernel-default-x86_64 --query Parameter.Value --output text)
$ aws ec2 create-launch-template --launch-template-name academy-web-lt --launch-template-data \
    "{\"ImageId\":\"$AMI\",\"InstanceType\":\"t3.micro\",\"SecurityGroupIds\":[\"$WEBSG\"],\"UserData\":\"$(base64 -w0 user-data.sh)\",\"MetadataOptions\":{\"HttpTokens\":\"required\"}}"
$ TG=$(aws elbv2 create-target-group --name academy-web-tg --protocol HTTP --port 80 --vpc-id $V --health-check-path / --query 'TargetGroups[0].TargetGroupArn' --output text)
$ ALB=$(aws elbv2 create-load-balancer --name academy-alb --type application --subnets $SUBS --security-groups $ALBSG --query 'LoadBalancers[0].LoadBalancerArn' --output text)
$ aws elbv2 create-listener --load-balancer-arn $ALB --protocol HTTP --port 80 --default-actions Type=forward,TargetGroupArn=$TG
$ aws autoscaling create-auto-scaling-group --auto-scaling-group-name academy-asg --launch-template LaunchTemplateName=academy-web-lt,Version='$Latest' \
    --min-size 2 --max-size 4 --desired-capacity 2 --vpc-zone-identifier "$(echo $SUBS | tr ' ' ,)" \
    --target-group-arns $TG --health-check-type ELB --health-check-grace-period 60
$ DNS=$(aws elbv2 describe-load-balancers --load-balancer-arns $ALB --query 'LoadBalancers[0].DNSName' --output text)
$ for i in 1 2 3 4; do curl -s http://$DNS; done            # after a few minutes: both AZs answer
$ aws ec2 terminate-instances --instance-ids <one of the group's instances>
$ aws autoscaling describe-scaling-activities --auto-scaling-group-name academy-asg --max-items 2 --query 'Activities[].Description'

aws ec2 authorize-security-group-ingress --group-id $ALBSG --protocol tcp --port 80 --cidr 0.0.0.0/0

awsthe AWS command line

for i in 1 2 3 4; do curl -s http://$DNS; done

forloop over a list
dostart of what the loop runs each time
curlmake an HTTP request
-ssilent: no progress bar
doneend of the loop

aws autoscaling describe-scaling-activities --auto-scaling-group-name academy-asg --max-items 2 --query 'Activities[].Description'

--querypick fields from the JSON response

Try it yourself

0 of 6 steps done

Run each task in your own account or terminal, then tick it off.

Create a launch template with the nginx user data from lesson 3.
Create an Auto Scaling group of 2 instances across two public subnets, with ELB health checks.
Put an Application Load Balancer in front and curl its DNS name a few times. Which hosts answer?
Terminate one instance by hand and watch the Activity tab.
Add a target tracking policy at 50% CPU and read what it will do.
Delete the group, load balancer, target group, launch template and security groups.

Knowledge check

Quick check

Question 1 of 5

What does an Auto Scaling group do when one of its instances fails its health check?

Go deeper

Official reading for this lesson

Or get every quiz answer right and it completes itself.