SOA-C03 · Domain 2 of 5

SOA-C03 Domain 2: Reliability and Business Continuity

This domain covers keeping AWS workloads available and recoverable: scaling with Auto Scaling and load balancing, building highly available multi-AZ designs, and implementing backup, restore and disaster-recovery strategies.

Exam weight
22%
Official task statements
3
Estimated questions estimate
~11
Suggested study time estimate
20–30 hrs

About 10 min read Reviewed October 2026

Overview

Reliability and Business Continuity covers operating workloads that stay available and recover from failure, from the operator's standpoint. It has three task statements. Scalability and elasticity covers EC2 Auto Scaling, scaling policies, load balancing and scaling of managed services. Highly available and resilient environments cover multi-Availability-Zone designs, Elastic Load Balancing health checks, RDS Multi-AZ and read replicas, and Route 53 failover. Backup and restore covers AWS Backup, snapshots, point-in-time recovery and disaster-recovery strategies such as backup and restore, pilot light, warm standby and active-active.

At 22%, this is a top-weighted domain. Unlike the architect exam, it focuses on operating and verifying these mechanisms: configuring health checks correctly, testing restores, and choosing a recovery strategy for stated objectives. It builds on the monitoring from Domain 1, which detects the failures this domain recovers from.

How it's tested

SOA-C03 questions describe an availability, scaling or recovery situation and ask how to configure or fix the mechanism. This domain rewards operational correctness: the right health check, scaling policy or recovery strategy for the stated objective. Read the objective carefully before choosing.

  • Scale correctly

    A load pattern is described, and you choose the Auto Scaling policy or load-balancing configuration that handles it.

  • Remove single points of failure

    An availability requirement is described, and you choose multi-AZ, ELB health checks or Route 53 failover.

  • Match the recovery strategy

    RTO and RPO are given, and you choose the backup or disaster-recovery approach that meets them.

Reasoning it through

A database must survive the loss of an Availability Zone with automatic failover, and separately the application needs to scale read traffic. An RDS read replica scales reads but does not provide automatic failover, while RDS Multi-AZ maintains a standby in another zone with automatic failover but does not scale reads. The correct operational answer uses Multi-AZ for availability and read replicas for read scaling, matching each mechanism to its purpose.

Task statements 3

Official task statements from the Amazon Web Services outline, in outline order.

2.1 Implement scalability and elasticityEC2 Auto Scaling, scaling policies and load balancing.
What it means
This covers EC2 Auto Scaling groups, launch templates, target tracking and step scaling policies, scheduled scaling, Elastic Load Balancing, and scaling of managed and serverless services to match capacity to demand.
What to understand
Auto Scaling adds and removes instances based on policies tied to metrics, and target tracking keeps a metric at a set value. Load balancers distribute traffic across healthy targets. Correct scaling requires sensible metrics, cooldowns and health checks so capacity matches demand without flapping.
How it shows up
A load pattern is described, and you must choose the Auto Scaling policy or load-balancing configuration.
Common confusion
Confusing target tracking with step scaling; target tracking maintains a metric value, while step scaling adjusts capacity in defined steps.
2.2 Implement highly available and resilient environmentsMulti-AZ designs, ELB health checks, RDS Multi-AZ and read replicas, Route 53 failover.
What it means
This covers deploying across multiple Availability Zones, configuring Elastic Load Balancing health checks, RDS Multi-AZ for failover and read replicas for read scaling, and Route 53 health checks and failover routing to maintain availability.
What to understand
Multi-AZ designs survive the loss of a zone. ELB health checks route traffic only to healthy targets. RDS Multi-AZ provides a standby with automatic failover, distinct from read replicas, which scale reads. Route 53 failover redirects DNS when an endpoint is unhealthy.
How it shows up
An availability requirement is described, and you must choose multi-AZ, the right RDS feature, health checks or Route 53 failover.
Common confusion
Confusing RDS Multi-AZ with a read replica; Multi-AZ is for failover, while read replicas are for read scaling.
2.3 Implement backup and restore strategiesAWS Backup, snapshots, point-in-time recovery and DR strategies.
What it means
This covers AWS Backup for centralized backups, EBS and RDS snapshots, point-in-time recovery, cross-Region and cross-account backup, and disaster-recovery strategies from backup and restore through pilot light and warm standby to active-active, chosen by RTO and RPO.
What to understand
Backups must be tested by restoring, not just taken. AWS Backup centralizes backup policies across services. Disaster-recovery strategies trade cost against recovery speed: backup and restore is cheapest with the highest RTO, active-active is costliest with near-zero RTO. The objectives decide the strategy.
How it shows up
A recovery requirement with RTO and RPO is described, and you must choose the backup or DR strategy.
Common confusion
Over-provisioning an active-active strategy for a loose RTO that backup and restore would meet more cheaply.

High-yield concepts

The terms, rules and formulas this domain turns into questions most often.

  • EC2 Auto Scaling

    Adds and removes instances based on policies and metrics.

    Remember
    Keeps capacity matched to demand.
    Tested as
    Scaling questions.
    Trap
    Omitting health checks or cooldowns.

  • Target tracking vs. step scaling

    Target tracking maintains a metric value; step scaling adjusts in steps.

    Remember
    Choose by how precisely you must hold a metric.
    Tested as
    Scaling questions.
    Trap
    Confusing the two policies.

  • Elastic Load Balancing health checks

    Route traffic only to healthy targets.

    Remember
    Misconfigured checks cause outages or flapping.
    Tested as
    Availability questions.
    Trap
    Setting health checks too aggressively or loosely.

  • RDS Multi-AZ

    A standby replica in another zone with automatic failover.

    Remember
    For availability, not read scaling.
    Tested as
    Database availability questions.
    Trap
    Confusing it with a read replica.

  • RDS read replica

    A copy that scales read traffic.

    Remember
    For read performance, not failover.
    Tested as
    Database scaling questions.
    Trap
    Using it for high availability.

  • Route 53 failover

    DNS health checks and failover routing.

    Remember
    Redirects traffic when an endpoint is unhealthy.
    Tested as
    Failover questions.
    Trap
    Ignoring DNS TTL effects on failover speed.

  • AWS Backup

    Centralized backup policies across services.

    Remember
    It standardizes and schedules backups.
    Tested as
    Backup questions.
    Trap
    Managing backups per service manually.

  • Point-in-time recovery

    Restoring data to a specific earlier moment.

    Remember
    Limits data loss to the recovery point.
    Tested as
    Recovery questions.
    Trap
    Assuming only full snapshots exist.

  • DR strategies

    Backup and restore, pilot light, warm standby, active-active.

    Remember
    Lower RTO and RPO cost more.
    Tested as
    DR questions.
    Trap
    Over-provisioning for a loose objective.

  • Testing restores

    Verifying backups by actually restoring them.

    Remember
    An untested backup may not work.
    Tested as
    Reliability questions.
    Trap
    Taking backups without testing restores.

Side by side

Concepts candidates mix up, compared on the point that decides the answer.

RDS Multi-AZ vs. read replica

ConceptMeaningKey differenceCommon trap
Multi-AZ Standby with automatic failover For availability Using it to scale reads
Read replica Copy that serves read traffic For read scaling Using it for failover

Target tracking vs. step vs. scheduled scaling

ConceptMeaningKey differenceCommon trap
Target tracking Hold a metric at a set value Simple, self-adjusting Confusing it with step scaling
Step scaling Adjust capacity in defined steps Fine control by threshold Overcomplicating simple cases
Scheduled scaling Scale at known times For predictable patterns Using it for unpredictable load

DR strategies by RTO and RPO

ConceptMeaningKey differenceCommon trap
Backup and restore Restore from backups Lowest cost, highest RTO Using it where RTO is short
Pilot light / warm standby Minimal or scaled-down standby Balances cost and speed Confusing the two scales
Active-active Full running in another Region Near-zero RTO, highest cost Over-provisioning for loose RTO

Exam traps

Each trap is an answer that looks right. Learn the clue that should change your mind.

  1. Read replica for failover

    What candidates pick
    Use an RDS read replica to provide automatic database failover.
    Why it looks right
    A replica is another copy.
    Why it's wrong
    Read replicas scale reads; RDS Multi-AZ provides a standby with automatic failover.

    Clue to watch for The requirement is automatic failover on zone loss.

  2. Multi-AZ for read scaling

    What candidates pick
    Add RDS Multi-AZ to handle heavy read traffic.
    Why it looks right
    Multi-AZ adds another instance.
    Why it's wrong
    The Multi-AZ standby does not serve reads; read replicas scale reads.

    Clue to watch for The need is to offload read queries.

  3. Active-active for loose RTO

    What candidates pick
    Build active-active multi-Region for a 24-hour RTO.
    Why it looks right
    It recovers fastest.
    Why it's wrong
    A generous RTO can be met with backup and restore or pilot light at far lower cost.

    Clue to watch for The RTO and RPO are generous and cost matters.

  4. Untested backups

    What candidates pick
    Rely on backups that have never been restored.
    Why it looks right
    Backups are being taken.
    Why it's wrong
    An untested backup may fail to restore; restores must be verified.

    Clue to watch for The question stresses recoverability or a real recovery event.

  5. Aggressive health checks

    What candidates pick
    Set a very short health-check interval and low threshold.
    Why it looks right
    It detects failure fast.
    Why it's wrong
    Overly aggressive checks cause false failures and flapping; checks must be tuned.

    Clue to watch for Instances cycle in and out of service unexpectedly.

Answer framework

For any Domain 2 scenario, use these steps to find the answer SOA-C03 expects.

  1. Identify whether the need is scaling, availability or recovery, since each has its own mechanisms.
  2. For scaling, choose the Auto Scaling policy that fits the load and configure health checks and cooldowns so capacity tracks demand smoothly.
  3. For availability, use multi-AZ, tuned ELB health checks and Route 53 failover, and choose RDS Multi-AZ for failover versus read replicas for read scaling.
  4. For recovery, read the RTO and RPO and pick the least costly strategy, from backup and restore to active-active, that meets them.
  5. Verify the design by testing restores and failover rather than assuming they work, and document the result.

Practice question

An application's database must automatically fail over to another Availability Zone if its primary becomes unavailable, and separately the application needs to offload heavy read traffic. Which combination should the CloudOps engineer implement?

Want more like this? The SOA-C03 practice test has questions for every domain.

Checklist

0 of 10

Tick each one you can do without notes. Your ticks stay in this browser.

Study strategy

Because this domain is operational, study how each reliability mechanism behaves and how to verify it, not just what it is. Start with Auto Scaling and load balancing: learn the scaling policy types and how health checks and cooldowns keep capacity matched to demand without flapping.

Next, master the availability mechanisms, and fix firmly the RDS Multi-AZ (failover) versus read replica (read scaling) distinction, which the exam tests repeatedly. Add ELB and Route 53 health checks and failover. Then learn backup and disaster recovery: AWS Backup, point-in-time recovery, and the DR strategies in order of cost and recovery speed, matched to RTO and RPO.

Most importantly, build the habit of testing restores and failover, since the exam and real operations both punish untested recovery. Hands-on practice in a free-tier account makes these concrete. Afterwards, move to Domain 3, deployment, provisioning and automation. Run a real restore from AWS Backup and force an RDS Multi-AZ failover in a lab, then review missed practice questions by recovery objective.

FAQ

How much of SOA-C03 is Domain 2?

Reliability and Business Continuity is 22% of the scored content, tied for the largest domain, and increased in weight from the previous exam version.

What is the difference between RDS Multi-AZ and a read replica?

RDS Multi-AZ maintains a standby in another Availability Zone with automatic failover for availability. A read replica is a copy that scales read traffic and does not provide automatic failover.

How do I choose a disaster-recovery strategy?

Match it to the required RTO and RPO. Generous objectives suit backup and restore or pilot light, while near-zero objectives need warm standby or active-active, at higher cost.

Why test restores?

Backups that are never restored may be incomplete or corrupt. Testing restores verifies that recovery will actually work when needed.

What scaling policy should I use?

Target tracking to hold a metric at a value, step scaling for threshold-based adjustments, and scheduled scaling for predictable, known load patterns.

How does Route 53 failover work?

Route 53 health checks monitor endpoints and failover routing redirects DNS to a healthy endpoint, though DNS TTL affects how quickly clients see the change.

How does Domain 2 connect to the rest of the exam?

It uses the monitoring from Domain 1 to detect failures and keeps the workloads available that the deployment, security and networking domains operate.

Other Cloud Computing certifications

Where the same subject shows up in other exams. Cards marked "Same subject" open the matching domain.

Test your Domain 2 knowledge

You have covered scaling, high availability and backup and recovery. Test whether you can configure the right mechanism and match recovery strategies to objectives under exam conditions.

Take the SOA-C03 practice test Check your SOA-C03 readiness

Keep going