OnCallReadyStart learning

AWS Lesson 39.42 · AWS VI: Architecting on AWS & the exams · 11 min read

The SOA-C03 exam: AWS Certified CloudOps Engineer - Associate

In plain words

If the Solutions Architect exam is about designing a house, the CloudOps exam is about being the building manager at three in the morning: the alarm went off, the lift is stuck, the heating bill doubled. Each question describes a symptom, and you choose the first thing to check or the fix that keeps it from happening again - preferably one that runs by itself.

The SOA-C03 exam: AWS Certified CloudOps Engineer - Associate

The CloudOps Engineer - Associate (exam code SOA-C03) replaced the SysOps Administrator - Associate in 2025. Where the Solutions Architect exam asks "what would you build?", this one asks "it is running and something is wrong - what do you look at, and what do you change?": metrics and alarms, logs, automation, backups, patching, deployments, security tooling and network troubleshooting. It is the exam closest to an SRE's day, and the one this course's labs and incidents prepare you for most directly.

Need to know: the same format as the Solutions Architect exam - 65 questions (50 scored, 15 unscored), 130 minutes, multiple choice and multiple response, a scaled score of 100-1,000 with 720 to pass, compensatory scoring, valid three years. The old SysOps exam had hands-on exam labs; they were dropped in 2023 and SOA-C03 has none - but its questions are written by people who expect you to have done the work, and they read like incident tickets.

The five domains

DomainWeightTask statementsWhere this course teaches it
1 Monitoring, logging, analysis, remediation, and performance optimization22%1.1 metrics, alarms, filters (CloudWatch, the agent, Logs Insights, dashboards) · 1.2 identify and remediate (EventBridge, SSM Automation, Config) · 1.3 performance optimization (EC2, EBS, S3, RDS)AWS III (CloudWatch, Logs, CloudTrail), AWS IV (EventBridge, SSM, Config), AWS V (RDS, EFS), this chapter (Lambda, ECS metrics)
2 Reliability and business continuity22%2.1 scalability and elasticity (Auto Scaling, caching, database scaling) · 2.2 HA and resilience (ELB, Route 53 health checks, Multi-AZ) · 2.3 backup and restore (AWS Backup, snapshots, PITR, DR)AWS II (ELB, ASG), AWS V (Backup, RDS, Route 53, ElastiCache), this chapter (resilience, scaling)
3 Deployment, provisioning, and automation22%3.1 provision and maintain (CloudFormation, StackSets, AMIs, deployment strategies, Terraform) · 3.2 automate existing resources (SSM, Lambda, EventBridge, Step Functions)Terraform chapters, AWS IV (CloudFormation, SSM, Patch Manager), this chapter (ECS deployments, Step Functions)
4 Security and compliance16%4.1 security and compliance tools and policies (IAM, Organizations, Config, Security Hub, Trusted Advisor) · 4.2 protect data and infrastructure (KMS, ACM, Secrets Manager, GuardDuty, Inspector, Macie)AWS I (IAM, KMS), AWS III (Organizations), AWS IV (Config, findings, secrets, ACM)
5 Networking and content delivery18%5.1 VPC features and connectivity · 5.2 DNS and content delivery (Route 53, CloudFront) · 5.3 troubleshoot connectivity (flow logs, Reachability Analyzer, ELB / CloudFront logs)AWS II (VPC, NAT, endpoints, flow logs), AWS V (Route 53 policies, CloudFront, TGW, VPN)

How the questions work

Most questions are a symptom plus a constraint: "users report 504 errors since this morning", "the instance does not appear in Systems Manager", "the backup job fails", "the stack is in UPDATE_ROLLBACK_FAILED", with "with the least operational effort" or "automatically" at the end. The answer is usually the first diagnostic or the managed automation:

Symptom in the questionThink
an alarm should fire but stays INSUFFICIENT_DATAthe metric, its dimensions, the period, missing-data treatment
an instance is not managed by SSMinstance profile with AmazonSSMManagedInstanceCore, the agent, a route or endpoints to SSM
"automatically remediate" a non-compliant resourcean AWS Config rule with an SSM Automation remediation
"notify / react when X happens"an EventBridge rule (and the target's resource policy)
a queue backlog grows, the function is idlea concurrency cap (reserved concurrency or MaximumConcurrency)
502 / 504 behind API Gateway or an ALBthe integration's errors vs its timeout
an ECS task never startsstoppedReason: the execution role, a route to ECR, the image tag
private instances cannot reach S3 cheaplyan S3 gateway endpoint
a deployment must not take the service downrolling with minimum healthy 100%, circuit breaker, canary / linear (Lambda aliases, CodeDeploy)

Study plan with this course

  1. Do the incidents. SOA questions are incidents in miniature. Every AWS chapter has them, and the phone app has each as a decision drill (Cards › Incident drills) - replay the ones you missed.
  2. Know the console and CLI names of the operational settings: alarm states and missing data, Auto Scaling cooldowns and warm pools, CloudFormation stack statuses, Patch Manager baselines and patch groups, AWS Backup plans and vaults, Route 53 health checks, flow log fields.
  3. Weak spots: Progress › Certifications ranks the SOA task statements by readiness; Focus 10 min on the weakest, then re-run its lab.
  4. Mocks: quick (20 questions) to find gaps, full (65 questions, 130 minutes) when quick ones are steady above 75%.
  5. Official: the SOA-C03 exam guide, the free official practice question set on AWS Skill Builder, and the CloudWatch, Systems Manager and AWS Backup user guides' troubleshooting pages.

Mock drills in this chapter

Right after this lesson come three timed SOA-C03 mocks in the terminal, drawn from the course's SOA-C03 question bank (original scenario questions written against the AWS documentation, never exam dumps), with new questions and shuffled options every round:

  • SOA-C03 mock A - domains 1-2 (monitoring and remediation, reliability and continuity): 15 questions, 20 minutes;
  • SOA-C03 mock B - domains 3-5 (deployment and automation, security, networking): 15 questions, 20 minutes;
  • SOA-C03 full mock - every domain by its weight: 40 questions, 50 minutes.

Each one writes its questions to ~/oncall-lab/labs/aws/exam/ and shows them in the brief. Answer all of them in one report: the question number and the letter(s), report 1b 2ac 3d; a "Choose two" item needs both letters. The pass mark is 70%. The debrief is the answer key with the reasoning and the lesson behind each question; drill starts another round.

SAA, SOA or both - and in which order

The two Associate exams overlap by roughly a third (VPC, IAM, Auto Scaling, Multi-AZ, backups, CloudWatch). After this course:

  • SAA first if you are moving into design or platform work, or the job adverts you read ask for it - it is the more recognised one.
  • SOA first if you are on call, run production, or want the exam that rewards the incident work of AWS III-VI most.
  • Both: take the second within a few weeks of the first while the overlap is fresh; most of the extra study is the other exam's specialities (SOA: Systems Manager, Config, CloudFormation operations, monitoring details; SAA: data stores, DR and cost design, analytics).

CLF-C02 (see the previous lesson) is optional before either.

Exam day

The same rules as for SAA-C03: two minutes per question, a first pass with flags, nothing blank, "least effort / automatically / managed" usually beats a script you would write yourself, and the question's last sentence first. For SOA in particular: when a question gives an error message, it is the clue - the course's error explainer taught you most of them (why after a failed command).

On the exam: this lesson is about the exam itself - format (65 questions, 130 minutes, 720 of 1,000, no exam labs), the five domains and weights (22 / 22 / 22 / 16 / 18), symptom-first questions and the managed-automation answers.

In an interview: "What is the difference between monitoring and observability, in AWS terms?" - "Monitoring is the known questions: CloudWatch metrics and alarms on what we expect to break, like queue age, 5XX rates, throttles. Observability is being able to answer new questions from the outside: structured logs I can query with Logs Insights, traces, CloudTrail for what changed, metrics with the right dimensions. On AWS I'd want both, and alarms that page on symptoms users feel, with the logs and traces to find the cause."

You can now: describe the SOA-C03 format and that it has no labs, map its five domains to the course, read a symptom-first question for the first diagnostic or the managed fix, plan revision with the incidents, readiness view, cards and mocks, and choose the order of SAA, SOA and CLF.

Why it helps

SOA-C03, the CloudOps Engineer - Associate, is the AWS exam closest to an SRE's daily work: monitoring, remediation, backups, deployments, security tooling and network troubleshooting. Its questions read like incident tickets, which is exactly what this course's labs and incidents train. Knowing its format, its five domains and the symptom-first question style - and where each task statement is taught in the course - turns that practice into a pass.

Commands in this lesson

aws

FAQ

Does SOA-C03 still have exam labs?

No. The old SysOps Administrator exam had hands-on labs until they were removed in 2023, and SOA-C03 is multiple choice and multiple response only: 65 questions, 50 of them scored, 130 minutes, 720 of 1,000 to pass. The questions still assume you have done the work hands-on.

What are the five SOA-C03 domains?

Monitoring, logging, analysis, remediation and performance optimization (22%); reliability and business continuity (22%); deployment, provisioning and automation (22%); security and compliance (16%); and networking and content delivery (18%). Each domain has task statements that Progress › Certifications maps to the lessons and labs that teach them.

What do "automatically" and "least effort" usually mean in SOA questions?

A managed mechanism instead of a script: an EventBridge rule with an SSM Automation runbook, an AWS Config rule with automatic remediation, target tracking scaling, an AWS Backup plan, a CloudFormation StackSet. Answers that require someone to log in and run commands are usually wrong when the question asks for automation.

How do the incident drills on the phone help?

Each AWS incident in the course is also a decision drill in the phone app: a symptom, then the next best move at each step, with explanations for the good, acceptable and wrong choices. That is the same reasoning SOA-C03 questions ask for, practised in a few minutes at a time.

Should I take SAA or SOA first?

SAA-C03 first if you are moving towards design or platform roles, SOA-C03 first if you run production and are on call. They overlap by about a third - VPC, IAM, Auto Scaling, Multi-AZ, backups, CloudWatch - so taking the second within a few weeks of the first saves study time.

In an interview Mid

What is the difference between monitoring and observability, in AWS terms?

Monitoring answers the known questions: CloudWatch metrics and alarms on what we expect to break, such as queue age, 5XX rates and throttles. Observability lets me answer new questions from outside: structured logs I can query with Logs Insights, traces, CloudTrail for what changed, and metrics with useful dimensions. I want both, with alarms on symptoms users feel and enough logs to find the cause - the skill behind the symptom-first questions of the CloudOps exam.

Also asked: What would you check first when an EC2 instance does not appear in Systems Manager? · How would you automatically remediate a non-compliant resource? · What are the domains of the CloudOps Engineer exam?

Practise this lesson in the terminal Free, in your browser - a real Ubuntu terminal to try it in, with missions that check your work.