SCS-C03 AWS Security – Specialty
- Vendor
- AWS
- Level
- Specialty
- Questions
- 65
- Time
- 170 min
- Domains
- 6
- In the bank
- 455
Practice 455 questionsTake a timed mock examDay-by-day study plan
Guides for this exam
The exam at a glance
| Level | Specialty |
|---|---|
| Questions | 65 |
| Time | 170 minutes (about 2.6 minutes per question) |
| Passing mark | 750 on a scaled score of 100–1,000 |
Domains and weights
| Domain | Weight | Practice questions here |
|---|---|---|
| D1 · Detection | 16% | 70 |
| D2 · Incident Response | 14% | 61 |
| D3 · Infrastructure Security | 18% | 78 |
| D4 · Identity and Access Management | 20% | 88 |
| D5 · Data Protection | 18% | 77 |
| D6 · Security Foundations and Governance | 14% | 81 |
61 of the 455 practice questions (13%) ask for more than one answer.
D1 · Detection 16%
Detection questions start from something that should have been seen and was not, or was seen only in pieces: a quiet Region, an alarm that never fires, separate findings that belong to one intrusion, a rejected flow in a log record. The task is to explain the gap or name the feature that closes it. Correct answers depend on knowing where each service collects data and what each log can and cannot tell you.
What to watch for
- Check scope before configuration. Several detection services work per Region, so a Region where nothing was enabled produces no findings at all.
- An alert built on CloudWatch Logs can only react to events that reach its log group. Confirm the delivery path from the trail before blaming the filter.
- Flow log records show that traffic was rejected but not which control rejected it. Think about stateful versus stateless rules and about return traffic on ephemeral ports.
Tasks in the official exam guide (3)
D2 · Incident Response 14%
Incident response scenarios are set in the middle of an event or just before one: high-severity findings that need containment, evidence that must stay untouched, ransomware risk, a leaked access key. The decision is which step comes first and which mechanism makes it reliable. Automated, isolated and immutable controls beat answers that depend on someone reading reports or on backups an attacker could reach.
What to watch for
- Order matters with exposed credentials: stop further use and cut off sessions already issued before any cleanup, so that evidence is not lost.
- For evidence or backups that nobody may delete, compare retention modes carefully. Some lock settings can be lifted by privileged users and others cannot.
- Keep detection and investigation apart: one service raises a finding, while another helps responders explore the related entities and activity over time.
Tasks in the official exam guide (2)
D3 · Infrastructure Security 18%
Infrastructure security items are mostly about network paths and hardened compute. A stem shows a routing setup, a private service that consumers must reach, an inspection design or an image process, and asks what is missing or which design fits. Many wrong options would work in a simpler network, so trace each flow in both directions, including return traffic, before deciding.
What to watch for
- When one service must be offered privately to many consumer VPCs without peering, think of exposing it as an endpoint service and controlling which principals may connect.
- Stateful inspection needs symmetric routing. When Network Firewall sits behind a transit gateway or in front of a public load balancer, check that both directions of a flow pass the same endpoint.
- Hardened instances come from a repeatable image pipeline, and private subnet traffic to S3 or DynamoDB usually does not need a NAT gateway at all.
Tasks in the official exam guide (3)
D4 · Identity and Access Management 20%
Identity questions often describe access that does not behave as expected or does not scale: permission sets that look stale, a policy count that grows with every project, servers outside AWS holding long-term keys, app users who need their own scoped access, or trust policies multiplying across clusters. The fix comes from understanding how each identity service issues and applies credentials. Before choosing, decide whether the problem is authentication or authorization.
What to watch for
- IAM Identity Center turns each permission set into a role in every assigned account. Edits made directly to those roles, or changes not yet provisioned, explain mismatched access.
- When the number of projects or clusters keeps growing, prefer designs that need no new policy or trust entry each time, such as tag matching or EKS Pod Identity.
- Long-term keys are rarely the answer. Servers outside AWS can trade certificates for temporary credentials, and app users can sign in and receive short-lived credentials scoped to them.
D5 · Data Protection 18%
Data protection questions first classify the data: moving over a network, stored at rest, or a secret or key in its own right. The scenario then adds a pressure (certificate renewal, KMS request cost, PII spread across many buckets, rotation without downtime, a second Region) and asks which feature handles it natively. Distractors often change how data is accessed instead of how it is protected.
What to watch for
- Ask who issued a certificate before assuming it renews itself. Certificates brought into ACM from outside follow a different lifecycle from those ACM issues.
- High KMS request volume from SSE-KMS has an S3-side fix that reduces calls to KMS; features that change how objects are read or filtered do not help.
- For decryption in a second Region, ask whether the same key material exists there. For rotation without downtime, consider how old credentials stay valid during the switch.
Tasks in the official exam guide (3)
D6 · Security Foundations and Governance 14%
Governance works at the organization level here: finding resources shared outside the organization, stopping vulnerable images before they reach a registry, keeping compliance evidence, weighing managed against custom guardrails, and removing root credentials from member accounts. Strong answers use a central, continuous mechanism. Watch for options that confuse AWS's own compliance reports with evidence about your own resources.
What to watch for
- Ask where in the lifecycle a control acts. Scanning an image after it is pushed is not the same as failing the build before the image gets there.
- Managed Control Tower guardrails reduce upkeep but are fixed rules, while custom SCPs cover company-specific needs. Questions comparing them often expect a combination rather than one winner.
- External access analysis depends on the zone of trust you set, and centralized root access management lets rare root-only tasks run without keeping root credentials in member accounts.
Tasks in the official exam guide (3)
A study order that works
The 28-day study plan for SCS-C03 takes the domains in this order. Each day has a short reading, key terms and a quiz from the bank.
- Days 1–4: D1 · Detection
- Days 5–8: D2 · Incident Response
- Days 9–13: D3 · Infrastructure Security
- Days 14–18: D4 · Identity and Access Management
- Days 19–23: D5 · Data Protection
- Days 24–27: D6 · Security Foundations and Governance
- Day 28: Final Practice Exam and Exam-Day Strategy
Recent changes to these questions
- : 3 answer choice sets changed, 3 answer keys corrected, 3 explanations rewritten
- : 60 questions added, 2 explanations rewritten
- : 5 answer choice sets changed, 5 answer keys corrected, 5 explanations rewritten
- : 1 explanation rewritten
- : 21 explanations rewritten
About these questions
Every question is original, written from the public exam guide and checked against the vendor's own documentation. None comes from a real exam. Each one has an explanation of why the answer is right and why the other options are wrong. How the questions are written.