No Docs Map — Closed Book
This page has no docs map to give you, because DOP-C02 gives you nothing to map. Some performance-based cloud-native certifications hand candidates a narrow allowlist of official documentation — a handful of sites, nothing else, but at least something — that they're permitted to open mid-exam. AWS's certifications, DOP-C02 included, are not that kind of exam. They are entirely closed-book: no AWS documentation, no console, no notes, no second screen, nothing installed locally to fall back on. Whatever is already in your head when you read a question is the entire universe of information available to answer it. That single fact should reshape how you study for the six domains covered under Certifications, and it's worth internalizing before you build a study plan around the wrong assumption.
Picture two closed-book tests. In the first, the teacher says you may bring one approved dictionary and nothing else — so part of studying is learning where the tricky word sits on the page. In the second, the teacher takes the dictionary away too. Suddenly "I'll know where to look it up" stops being a study plan, because there's no "where" left to look. DOP-C02 is the second test. Every word has to already be in your head before you walk in.
Why this page exists instead of a docs map
☺ Like you're 10: Some tests let you bring one approved book in with you. This one doesn't let you bring anything, so there's no book left to draw a map of.
AWS states plainly, across its certification policies, that its exams are closed-book. During the sitting you may not use notes, scratch paper you brought yourself, a phone, a second device, or any outside reference material — and the Candidate Code of Conduct you agree to at registration explicitly names unauthorized AI, machine-learning, and automated tools as prohibited too, not just old-fashioned cheat sheets. At a physical testing center you'll typically be asked to turn your pockets out at check-in to confirm they're empty. The one detail that does vary is what you're allowed to write on: testing centers commonly provide a physical writing surface such as a small laminated whiteboard, while OnVUE's online-proctored sitting instead gives you a built-in digital whiteboard tool inside the exam software — neither is a substitute for outside documentation, and the exact mechanics can differ by exam and by delivery method, so don't assume anything about your own sitting without checking your registration confirmation directly.
Everything on this page describing what is and isn't permitted during an AWS Certification exam was checked directly against AWS's own Candidate Code of Conduct and Certification Policies pages in August 2026. AWS revises exam-day policy language periodically, and specifics like whether a physical or digital whiteboard is provided can differ by testing center, region, or delivery method. Treat this page as "true as of this check," not as the authoritative source — read AWS's own General Policies and During Testing pages yourself in the days before you sit DOP-C02, and let those override anything written here.
↗ AWS's official "During Testing" policy page
What closed-book actually rules out
☺ Like you're 10: An allowlisted exam gives you a locked toolbox with a few tools in it. This exam gives you an empty toolbox — the only tools you get are the ones already in your hands.
It helps to see the two models side by side, because candidates coming from a hands-on, allowlisted certification often carry the wrong mental model into DOP-C02 without noticing. An allowlisted performance exam still trusts you to reason and look things up within limits — its docs map exists precisely because there's a real, bounded set of things you're permitted to open. DOP-C02 has no such set. The bar below on the right isn't a smaller allowlist; it's the absence of one.
Concretely, that means: no AWS documentation or console open in another tab or window, no personal notes or scratch paper you brought with you, no phone or second monitor, no talking through the problem out loud except to the proctor if something's genuinely wrong, and no AI or automated assistance of any kind — that last one is named explicitly in the Candidate Code of Conduct, not left implicit. The only writing surface is whatever the delivery method provides (a testing center's physical whiteboard, or OnVUE's digital one), and it exists for working through a problem on the page, not for smuggling in anything you wrote before the clock started.
Which facts have to live in your head
☺ Like you're 10: Reading a service's docs feels like knowing it. The exam only rewards recalling it cold, with nothing open to lean on.
Once looking things up is off the table entirely, three categories of fact stop being "nice to know" and become the actual exam. None of them require memorizing an entire service's documentation — they require memorizing the handful of details that a scenario question is actually testing, cold.
| Category | Why closed-book makes it bite | Example |
|---|---|---|
| Confusable service names | Two similarly named services doing genuinely different jobs; the wrong name is the wrong answer, and there's no doc tab to sanity-check it against. | CodeBuild compiles and tests source into an artifact; CodeDeploy takes that already-built artifact and deploys it to EC2, on-prem, Lambda, or ECS; CodePipeline is the thing that chains those stages together; CodeArtifact just stores and versions packages. Four names, four different jobs. |
| Default behaviors | Scenario questions love "what happens if you configure nothing extra" — and a default you can't check is a default you either know or guess. | CodeDeploy's classic EC2/on-prem deployment configuration defaults to CodeDeployDefault.OneAtATime; a new CloudWatch Logs log group defaults to Never Expire retention; S3 bucket versioning defaults to disabled and, once enabled, can only be suspended — never fully turned back off. |
| Numeric limits & thresholds | Quota questions are unanswerable by reasoning alone — a number either matches what you carry in your head or it doesn't. | Lambda memory runs 128 MB–10,240 MB, timeout tops out at 900 seconds (15 min), and a direct .zip deployment package is capped at 50 MB zipped / 250 MB unzipped. Systems Manager Parameter Store's Standard tier caps a parameter value at 4 KB; Advanced tier raises that to 8 KB and is what parameter policies require. |
| Named rollout mechanics | Deployment-strategy questions are easy to blur under time pressure because the names sound similar and the percentages matter. | CodeDeploy's built-in Lambda/ECS traffic-shifting configurations are named and fixed — Canary10Percent5Minutes ships 10% of traffic, waits five minutes, then cuts over the rest; Linear10PercentEvery1Minute shifts in 10% steps once a minute. You pick the named configuration that matches the requirement; you don't write your own canary math. |
Don't try to memorize a service's entire spec sheet — that's the read-it-again trap, and it produces recognition, not recall. Memorize the shape: which service owns which job, what the default is when a question says nothing extra, and the one or two numbers per service that keep reappearing across the heavily weighted domains. See the DOP-C02 exam guide for exactly how much of the exam D1, D2, and D6 are worth — that weighting is where the highest-value memorization lives.
The specific figures in the table above — memory ranges, timeout ceilings, parameter-value caps — are accurate as of this writing, but AWS revises service quotas over time, sometimes upward, without much fanfare. Cross-check anything you intend to bank marks on against the AWS Service & Command Reference and, closer to exam day, against AWS's own Service Quotas console or the service's FAQ page. A stale number memorized with total confidence is worse than an uncertain one, because you won't know to doubt it.
Study strategy for a fully closed-book exam
☺ Like you're 10: On an allowlisted test, "know where to look" is a real study skill worth practicing. On this one, that skill is worth exactly zero — so practice the other one instead: instant recall.
If you've studied for an allowlisted, hands-on certification before, drop the instinct it taught you. There, "I can find this fast under pressure" was a legitimate, practiced skill — knowing which doc page has the field you need, and how quickly you can get to it, was part of what the exam rewarded. On DOP-C02, that skill has no exam to apply to. The only thing that transfers is recall speed and recall accuracy under time pressure, which means your study sessions should be built around producing that, not around producing familiarity.
- Active recall over passive reading. Rereading a service's docs feels like progress and mostly measures whether you can recognize the answer when it's already in front of you — which the exam never lets you do. Closing the material and writing what you remember is what actually builds the thing DOP-C02 tests. Flashcards and Know It Cold are built around exactly this loop.
- Study with the same constraint the exam enforces. Practicing with a documentation tab open "just to double-check" trains a skill you won't have on exam day and inflates your sense of readiness. Do full read-throughs with references open if you need to build understanding first — but do every drill and every practice question with everything closed, the same way you'll sit the real thing.
- Full-length timed mock exams, closed-book, no exceptions. Working Mock Exam · Set 1 through Set 5 under a genuine 180-minute clock with nothing open is the only way to find out whether your recall actually holds up once fatigue and time pressure are real variables, not once you've reread the answer key.
- When recall genuinely fails mid-question, that's a process, not a doc lookup. There's no tab to alt-tab to, so the fallback isn't "go look it up" — it's eliminating options you're confident are wrong and narrowing from there. Answer Triage & Elimination covers that process in full.
Pick one domain — start with D1, SDLC Automation, since it's the heaviest-weighted. Close every browser tab and put your phone out of reach. Set a ten-minute timer and write, from memory only: every CodeCommit/CodeBuild/CodeDeploy/CodePipeline/CodeArtifact name paired with the one job each actually does; CodeDeploy's default deployment configuration name for an in-place EC2 deployment; and the two named Lambda/ECS traffic-shifting configurations along with what percentage each ships first. Then, and only then, open the Service & Command Reference and grade yourself. Anything you got wrong — or hesitated on — goes on a flashcard, not back on a re-read list.
Where this fits in the rest of exam prep
☺ Like you're 10: This page is the warning label. The other exam-prep pages are where you actually build the memory this page says you need.
This page's whole job is to reset one assumption before it costs you marks: there is no docs map for DOP-C02 because there is nothing to map, and every minute you'd otherwise spend planning "where to look things up" is better spent building recall instead. From here, the DOP-C02 exam guide covers the format and domain weights that tell you where to point that recall-building effort, and the study plan turns those weights into an actual week-by-week schedule. The Service & Command Reference is the material to drill against, Know It Cold is where the drilling itself lives, and Answer Triage & Elimination is the process for the moment recall still fails you mid-question despite all of it. When you're ready to find out whether it actually stuck, Mock Exam · Set 1 is the place to start — closed-book, timed, no exceptions.
Foxy: Wait, so if I go blank on the CodeDeploy default, I can't even peek at a doc tab for ten seconds?
Gizmo: Just alt-tab real quick, nobody's watching. 🤑
Ellie the Elephant: Everybody's watching, Gizmo — proctor, webcam, and the Code of Conduct you agreed to at registration. And there's no tab to alt-tab to anyway. The whole exam runs closed-book, by design, not by luck.
Remy the Rabbit: Which is why I don't "look things up." Ask me OneAtATime versus AllAtOnce and I answer before you finish the question — because I drilled it blank-page, same as this exam expects.
Timmy the Turtle: And if Remy's memory ever does blank mid-question, that's not a doc lookup — that's triage. Eliminate what you know is wrong first. Same discipline as verifying a deployment: work the process, don't panic.
Professor Owl: Some exams teach you where to look. This one only tests what's already in your head. Study accordingly, and the closed-book part stops being scary — it's just the format.
That's the whole point of this page: no allowlist, no docs map, no fallback — so build the recall now, while you still have a browser tab open to build it with. Next stop, Know It Cold to start the drilling, or back to the study plan to see where this fits in your schedule.
1. Name three things explicitly prohibited during an AWS Certification exam sitting, beyond generic "no cheating." 2. What's the actual difference between an allowlisted cloud-native exam's docs map and DOP-C02's closed-book format — why is DOP-C02 not just "a smaller allowlist"? 3. Give an example each of a confusable service-name pair, a default behavior, and a numeric limit that DOP-C02 could test without warning. 4. Why does practicing with a documentation tab open "just in case" actively hurt your exam readiness, according to this page? 5. When your memory genuinely fails mid-question on exam day, what should you do instead of trying to look something up?
Check your answers
- Any three of: personal notes or scratch paper brought from outside, a phone or second device, outside reference material or documentation, unauthorized AI/automated tools, or talking through the problem out loud to anyone but the proctor.
- An allowlisted exam still permits a real, bounded set of resources you're allowed to open mid-exam — hence a "docs map" of that set is genuinely useful. DOP-C02 permits none of that set; it isn't a smaller version of the same model, it's the absence of the model entirely. Whatever isn't already in your head when you read the question isn't available at all.
- Any reasonable example, e.g.: CodeBuild vs. CodeDeploy vs. CodePipeline (build vs. deploy vs. orchestrate); CodeDeploy's default in-place EC2 deployment configuration,
CodeDeployDefault.OneAtATime; Lambda's memory range of 128 MB–10,240 MB or its 900-second timeout ceiling. - It trains a skill — "look it up fast" — that has no exam to apply to, and it inflates your sense of readiness by letting you succeed on a practice question in a way you can't replicate on exam day, so your practice scores stop predicting your real performance.
- Fall back on triage, not lookup: eliminate the answer options you're confident are wrong first and narrow down from there, the same disciplined elimination process covered in Answer Triage & Elimination — there is no tab to check instead.