Validation Quiz - MODULE 8
This 30-question quiz allows you to validate your knowledge on high availability and load balancing before moving on to the practical lab.
Question 1
What does HA (High Availability) mean?
A) Hardware Architecture
B) High Availability - the ability of a system to remain operational despite failures
C) High Access - fast data access
D) Host Allocation
See the answer
Answer: B
High Availability (HA) is the ability of a system to remain operational and accessible even in the event of a component failure. It is measured in availability percentage (e.g., 99.99% = less than 52 minutes of downtime per year).
Question 2
What does "99.99%" availability (four nines) mean?
A) 99.99% of users are satisfied
B) Less than 52 minutes of downtime per year
C) 99.99% of requests succeed
D) The system is 99.99% cheaper
See the answer
Answer: B
99.99% availability means the system can be unavailable for a maximum of 52.56 minutes per year (0.01% of the year). 99.9% allows 8.76 hours, 99.999% only 5.26 minutes. The higher the number of "9s," the more demanding and costly it is.
Question 3
What is the main strategy to achieve high availability?
A) Use more powerful servers
B) Redundancy - deploy in multiple availability zones
C) Increase memory
D) Use SSD drives
See the answer
Answer: B
Redundancy is the key to HA. By deploying identical resources in multiple AZs (or regions), a failure in one zone does not affect overall availability. Traffic is redirected to healthy resources.
Question 4
How many minimum availability zones for a high availability architecture?
A) 1
B) 2
C) 3
D) 5
See the answer
Answer: B
At least 2 AZs for high availability. If one AZ fails, the other takes over. AWS often recommends 3 AZs for more resilience, especially for critical workloads. Each AWS region has at least 3 AZs.
Question 5
What is a Single Point of Failure (SPOF)?
A) A single component whose failure causes the entire system to be unavailable
B) A very powerful server
C) A security control point
D) A performance measure
See the answer
Answer: A
A SPOF is a non-redundant component whose failure causes the complete service outage. The goal of HA is to eliminate SPOFs by duplicating each critical component. Example of a SPOF: a single database instance without a replica.
Question 6
What are the three types of AWS Load Balancers?
A) Small, Medium, Large
B) ALB, NLB, GLB (Application, Network, Gateway)
C) HTTP, TCP, UDP
D) Public, Private, Hybrid
See the answer
Answer: B
AWS offers: Application Load Balancer (ALB) level 7 for HTTP/HTTPS, Network Load Balancer (NLB) level 4 for high-performance TCP/UDP, and Gateway Load Balancer (GLB) for virtual appliances (firewalls, IDS). The Classic Load Balancer is deprecated.
Question 7
Which Load Balancer to choose for a standard web application?
A) Network Load Balancer
B) Application Load Balancer
C) Gateway Load Balancer
D) Classic Load Balancer
See the answer
Answer: B
The ALB is ideal for HTTP/HTTPS web applications. It supports routing based on URL, headers, and cookies. It can also perform SSL termination and integrate WAF. The NLB is for cases requiring extreme performance or non-HTTP protocols.
Question 8
Which Load Balancer offers the lowest latency?
A) Application Load Balancer
B) Network Load Balancer
C) Gateway Load Balancer
D) They all have the same latency
See the answer
Answer: B
The NLB operates at level 4 (transport) and offers very low latency (microseconds vs milliseconds for the ALB). It can handle millions of requests per second. Ideal for real-time applications, gaming, IoT.
Question 9
What is a Target Group?
A) A group of target users
B) A set of targets (instances, IPs, Lambda) to which the LB sends traffic
C) A group of Security Groups
D) A group of metrics
See the answer
Answer: B
A Target Group groups the targets that receive traffic from the Load Balancer. Targets can be EC2 instances, IP addresses, or Lambda functions. Each Target Group has its own health check settings.
Question 10
What is a Health Check?
A) A medical examination
B) A periodic check of the health status of targets by the Load Balancer
C) A security audit
D) A performance test
See the answer
Answer: B
The Health Check is a periodic check sent by the LB to the targets. If a target does not respond correctly (bad HTTP code, timeout), it is marked "unhealthy" and no longer receives traffic until it recovers.
Question 11
What parameters configure a Health Check?
A) Only the URL
B) Protocol, port, path, interval, healthy/unhealthy thresholds, timeout
C) Only the port
D) Only the timeout
See the answer
Answer: B
Parameters include: protocol (HTTP, HTTPS, TCP), port, path (/health), interval between checks, number of successful checks to be healthy, number of failures to be unhealthy, timeout, and expected response codes.
Question 12
A target becomes "unhealthy" after how many consecutive failures by default?
A) 1
B) 2
C) 5
D) 10
See the answer
Answer: B
By default, after 2 consecutive failures (unhealthy threshold = 2), a target is marked unhealthy. Similarly, it takes 5 consecutive successes (healthy threshold = 5) by default to become healthy again. These values are configurable.
Question 13
What happens if all targets in a Target Group are unhealthy?
A) The LB returns a 503 error to all clients
B) The LB can be configured to send traffic anyway (fail-open)
C) The LB stops
D) A and B depending on the configuration
See the answer
Answer: D
By default, if all targets are unhealthy, the LB returns errors. But "fail-open" can be enabled to continue distributing traffic to all targets even if unhealthy, hoping they sometimes respond. Use with caution.
Question 14
What is Cross-Zone Load Balancing?
A) Load balancing between regions
B) Distributing traffic evenly among all targets across all AZs
C) Geographic routing
D) Data replication
See the answer
Answer: B
Cross-Zone Load Balancing distributes traffic evenly among all registered targets, regardless of their AZ. Without this option, each LB node distributes only to targets in its own AZ, which can create imbalances.
Question 15
Is Cross-Zone Load Balancing enabled by default for ALB?
A) No
B) Yes
C) Depends on the region
D) Only for large volumes
See the answer
Answer: B
For ALB, Cross-Zone Load Balancing is enabled by default and cannot be disabled. For NLB, it is disabled by default (for performance reasons) and can be enabled per Target Group.
Question 16
What is SSL termination at the Load Balancer level?
A) Stop using SSL
B) The LB decrypts HTTPS traffic and communicates in HTTP with backends
C) Renew certificates
D) Block SSL connections
See the answer
Answer: B
SSL termination means the LB handles HTTPS encryption/decryption. Clients communicate in HTTPS with the LB, and the LB communicates in HTTP with backend instances (in the secure private network). Simplifies certificate management.
Question 17
Where to store SSL certificates for an ALB?
A) On EC2 instances only
B) In AWS Certificate Manager (ACM)
C) In S3
D) In the application code
See the answer
Answer: B
ACM (AWS Certificate Manager) is the recommended service for storing and managing SSL/TLS certificates. ACM certificates are free for integrated services (ALB, CloudFront) and are automatically renewed.
Question 18
What is a Listener on a Load Balancer?
A) A logging tool
B) A process that listens on a specific port/protocol and routes according to rules
C) A security audit
D) A type of instance
See the answer
Answer: B
A Listener listens for connections on a port and protocol (e.g., HTTPS 443). When a request arrives, it evaluates it against configured rules and routes it to the appropriate Target Group. A LB can have multiple Listeners.
Question 19
How to automatically redirect HTTP to HTTPS with an ALB?
A) It is not possible
B) Create a redirection rule on the HTTP Listener (port 80)
C) Modify the application code
D) Use Route 53
See the answer
Answer: B
On the HTTP Listener (80), create a redirection action to HTTPS (301 or 302). All HTTP requests will be automatically redirected to HTTPS. It's a security best practice.
Question 20
What is path-based routing?
A) Routing based on the source IP address
B) Route to different Target Groups based on the requested URL
C) Geographic routing
D) Time-based routing
See the answer
Answer: B
Path-based routing allows routing /api/* to an API Target Group, /images/* to another, /app/* to a third. Very useful for microservices architectures or CDNs.
Question 21
In which layers should the ALB be deployed?
A) Private subnets only
B) Public subnets (at least 2 AZs)
C) A single AZ is enough
D) Anywhere
See the answer
Answer: B
An internet-facing ALB should be deployed in public subnets (with route to IGW) in at least 2 AZs for high availability. The LB creates nodes in each configured AZ.
Question 22
What is an "internal" vs "internet-facing" Load Balancer?
A) Same thing
B) Internal has a private IP, internet-facing has a public IP
C) Internal is faster
D) Internet-facing is free
See the answer
Answer: B
An internet-facing LB has public IP addresses and can receive traffic from the Internet. An internal LB has only private IPs and is used to distribute traffic between internal tiers (e.g., between web and application layer).
Question 23
Can a WAF be attached to an ALB?
A) No
B) Yes, AWS WAF can be attached to an ALB
C) Only to a NLB
D) Only to CloudFront
See the answer
Answer: B
AWS WAF (Web Application Firewall) can be attached to an ALB to filter malicious requests (SQL injection, XSS, etc.) before they reach applications. WAF also works with CloudFront and API Gateway.
Question 24
Which metric to monitor for a Load Balancer?
A) CPUUtilization
B) RequestCount, TargetResponseTime, HTTPCode_ELB_5XX, HealthyHostCount
C) DiskReadOps
D) NetworkIn only
See the answer
Answer: B
Key metrics: RequestCount (volume), TargetResponseTime (latency), HTTPCode_ELB_5XX (LB errors), HTTPCode_Target_5XX (backend errors), HealthyHostCount/UnHealthyHostCount (target health), ActiveConnectionCount.
Question 25
What does an increase in Load Balancer 5XX errors (ELB_5XX) mean?
A) Client-side issue
B) The LB has trouble reaching backends or all are unhealthy
C) DNS issue
D) Too much traffic
See the answer
Answer: B
ELB_5XX errors indicate a problem at the LB level: unable to connect to backends (all unhealthy, Security Group blocking, etc.). Target_5XX errors indicate that backends are responding with errors.
Question 26
How does the LB know which instance to send traffic to?
A) Randomly
B) Round-robin by default, or least connections, or sticky sessions
C) Always the first one
D) The closest geographically
See the answer
Answer: B
By default, the ALB uses round-robin (each request to the next target). You can enable least outstanding requests (to the target with the fewest ongoing requests) or sticky sessions (cookie to stay on the same target).
Question 27
What is Connection Draining (deregistration delay)?
A) Emptying logs
B) Allowing time for ongoing requests to complete before removing a target
C) Deleting inactive connections
D) Limiting the number of connections
See the answer
Answer: B
Connection Draining allows ongoing requests to complete gracefully when an instance is removed from the Target Group (scaling down, maintenance). Default is 300 seconds. New requests are no longer sent.
Question 28
A website has predictable load spikes. What strategy to adopt?
A) Oversize permanently
B) Pre-scaling with scheduled scaling policies
C) Nothing, the LB handles everything
D) Turn off servers outside peak times
See the answer
Answer: B
For predictable spikes (Black Friday, events), use scheduled scaling to increase capacity before the spike. The LB distributes traffic, but instances must exist. Reactive scaling may be too slow for sudden spikes.
Question 29
What is the difference between vertical and horizontal scalability?
A) None
B) Vertical = more powerful instance, Horizontal = more instances
C) Vertical = more instances, Horizontal = more power
D) Depends on the cloud provider
See the answer
Answer: B
Vertical scalability (scale up): increase resources of an instance (more CPU, RAM). Horizontal scalability (scale out): add more instances. The cloud favors horizontal as it allows better resilience and is almost unlimited.
Question 30
How to test the high availability of an architecture?
A) Wait for a real failure
B) Simulate failures: stop an instance, an AZ, check failover
C) Read the documentation
D) Trust AWS
See the answer
Answer: B
You must actively test HA by simulating failures: terminate an instance and check traffic failover, simulate an AZ failure, test health check behavior. AWS Fault Injection Simulator can help automate these tests (Chaos Engineering).
Result
Count your correct answers:
- 27-30: Excellent - You master high availability
- 22-26: Good - Some points to review
- 17-21: Average - Review the course before the lab
- Less than 17: Insufficient - Start the course from the beginning
Proceed to the HA Architecture Lab once you have achieved at least 22 correct answers.