AWS High Availability Design Patterns for the SAA-C03 Exam

High availability (HA) is a core concept on the AWS SAA-C03 exam. Questions often test your ability to architect resilient, fault-tolerant systems using services like Elastic Load Balancing (ELB), Auto Scaling groups (ASG), Amazon RDS Multi-AZ, Route 53 health checks, and Multi-Region strategies. This post breaks down the key patterns you need to know.

Multi-AZ vs Multi-Region: When to Use Which

Multi-AZ protects against an Availability Zone failure. It provides low-latency failover within a single region. Use for applications that require high availability but can tolerate a few seconds of failover time (e.g., RDS Multi-AZ, an active-passive setup with synchronous replication).

Multi-Region protects against a region-wide disaster. It introduces higher latency and complexity. Use only when your application must survive a full region outage (e.g., global user base, regulatory compliance). On the exam, if the question mentions “disaster recovery” or “region failure,” think Multi-Region. If it says “AZ failure” or “high availability,” Multi-AZ is usually sufficient.

Auto Scaling Groups and Elastic Load Balancing

ASG + ELB is the classic HA pattern for compute. The ASG maintains a desired number of EC2 instances across multiple AZs. The ELB distributes traffic to healthy instances.

Key details for the exam:

• ASG can span multiple AZs. Always specify at least two AZs for HA.
• ELB health checks (HTTP/HTTPS on port 80/443, or TCP on port 443) determine instance health. Unhealthy instances are replaced.
• ELB has a connection draining feature (deregistration delay) to complete in-flight requests before terminating instances.
• Use Application Load Balancer (ALB) for HTTP/HTTPS traffic; Network Load Balancer (NLB) for TCP/UDP with ultra-low latency.

Example scenario: A web app runs on EC2 behind an ALB. The ASG has min=2, max=10, desired=2 across us-east-1a and us-east-1b. If one AZ fails, the ASG launches a new instance in the remaining AZ, and the ALB routes traffic only to healthy instances.

RDS Multi-AZ for Database HA

Amazon RDS Multi-AZ provides automatic failover to a standby replica in a different AZ. Synchronous replication ensures no data loss. Failover typically takes 1-2 minutes.

Exam tips:

• RDS Multi-AZ is for HA only, not for read scaling. For read replicas, use Read Replicas (asynchronous).
• The standby is not accessible for reads/writes until failover.
• Supported engines: MySQL, PostgreSQL, Oracle, SQL Server, MariaDB.
• Failover is triggered by AZ failure, primary instance failure, or manual reboot with failover.

Example: An application uses RDS MySQL Multi-AZ. The primary DB in us-east-1a fails. Route 53 health check (if using custom DNS) or the RDS endpoint automatically redirects to the standby in us-east-1b. The application experiences a brief interruption but recovers without manual intervention.

Route 53 Health Checks and DNS Failover

Route 53 health checks monitor the health of your resources (e.g., ELB, EC2, endpoints). Combined with routing policies, they enable DNS-level failover.

Routing policies for HA:

• Failover routing: Primary and secondary records. If primary fails health check, traffic goes to secondary.
• Weighted routing: Distribute traffic across multiple resources; can be used for blue/green deployments.
• Latency routing: Route to the region with lowest latency, but not inherently HA.
• Geolocation/Geoproximity: For compliance or traffic distribution.

Exam tips:

• Health checks can monitor an endpoint (HTTP/HTTPS/TCP) or other health checks (calculated).
• Health check interval: 30 seconds (default) or 10 seconds (fast).
• For ELB, enable “Evaluate target health” on the ALB/NLB target group, and Route 53 health check against the ELB DNS name.

Example: You have an ALB in us-east-1 (primary) and another in us-west-2 (secondary). Route 53 failover routing: primary record with health check on the us-east-1 ALB. If the health check fails, traffic is routed to us-west-2.

What to Watch for on the Exam

1. Read the question carefully: Does it ask for “high availability” (Multi-AZ) or “disaster recovery” (Multi-Region)?
2. Cost vs. availability trade-offs: Multi-AZ is cheaper than Multi-Region. The exam may ask you to choose the most cost-effective HA solution.
3. Stateless vs. stateful: For stateless web tiers, use ASG + ELB. For stateful databases, use RDS Multi-AZ or DynamoDB Global Tables.
4. ELB types: ALB for HTTP/HTTPS (Layer 7), NLB for TCP/UDP (Layer 4), Classic Load Balancer (deprecated, avoid).
5. Health check depth: Route 53 health check can check an ELB, which checks its target group health. Understand cascading health checks.
6. Auto Scaling cool-downs: After a scaling activity, ASG waits for a cool-down period (default 300 seconds) before launching another instance.
7. RDS Multi-AZ failover time: Expect ~1-2 minutes. If the question requires <1 second failover, consider Aurora (which has faster failover) or DynamoDB.

Conclusion

Mastering these HA patterns is essential for the SAA-C03 exam. Remember: Multi-AZ for AZ failures, Multi-Region for region failures. Combine ASG + ELB for compute, RDS Multi-AZ for databases, and Route 53 health checks for DNS failover. Practice identifying the right pattern based on requirements like cost, latency, and recovery time objective (RTO).

To solidify your knowledge, try practice questions that ask you to design a fault-tolerant architecture. Focus on scenarios where an AZ goes down, and ensure your design survives without manual intervention.