The SREF Study Plan
Eight modules, forty questions, sixty minutes, closed book. The SRE Foundation is the shortest exam this course prepares you for, and that brevity is exactly what makes it easy to under-plan — read the eight blueprint modules once, in order, over a long weekend, and walk in hoping recognition turns into recall once the clock starts. This page is the corrective: a four-week plan that gives each of the eight blueprint modules its own suggested day count, pairs it with the course lesson or lessons that actually teach the idea in depth, tells you how to spend five distinct mock exams without wasting any of them, and closes with a short list of objective gates — not a feeling — that say you're actually ready to book a seat.
Imagine eight chapters to learn for a quiz that mixes questions from all eight, roughly evenly. Reading all eight chapters once, back to back, the night before feels like studying — it isn't the same as knowing each chapter cold with a clock running the next morning. This page is a calendar that gives each chapter its own number of days, tells you exactly which page of your existing textbook matches each quiz topic, and saves the last week for five practice quizzes instead of more reading — because the way you find out a chapter didn't stick is by being tested on it, not by rereading it a third time.
What you're actually studying for
☺ Like you're 10: Forty quick questions, four choices each, one hour, no notes — and, as best anyone can tell, all eight chapters get roughly the same number of questions.
The SRE Foundation is closed-book, single-best-answer multiple choice: 40 questions in 60 minutes, one correct option per question with no partial credit for a well-reasoned wrong answer, and 65% — 26 or more correct — to pass. There are no formal prerequisites. Full exam-day logistics, registration paths, cost, and delivery details live on The SREF Exam; this page's job is narrower and specific — sequencing the study time you have left, not the booking process.
One structural fact from that page shapes every week below: DevOps Institute does not publish a per-module question weighting for the 40-question paper, unlike the CNCF's hands-on exams elsewhere on this site, which publish exact domain percentages. This course's own five mock exams therefore assign exactly five questions to each of the eight modules — an explicit, stated heuristic, not a number DevOps Institute has confirmed in writing. The day counts in this plan lean on a different axis than exam weighting, and it's worth being precise about the difference: they're sized by how much course material each module actually pairs with, not by a guess about how many exam points it's worth. A module that takes three days to study isn't necessarily worth more marks than one that takes two — it just has more ground underneath it. Treat all eight as equally capable of costing you points, and let this plan's days reflect study load, not importance.
Question count, timing, the pass mark, and registration mechanics all change over time, and this page reflects the landscape described on The SREF Exam as of 2026. Confirm current specifics on DevOps Institute's and PeopleCert's own pages before you register — this plan's day counts and sequencing survive a minor format change; the exact numbers you'd verify against a vendor page might not.
Scope-wise, this plan assumes short, near-daily sessions — roughly 45–60 minutes a day, somewhere around 20–25 hours total across four weeks. That's deliberately light next to the 150-hour plan a hands-on, performance-based exam elsewhere on this site's sister course demands — a 60-minute closed-book paper earns a proportionate investment, not a smaller copy of a much bigger exam's plan.
Place yourself first — how many weeks you actually need
☺ Like you're 10: Some people already half-know this from doing the job; others are starting from zero. Six honest checks tell you whether four weeks is too long, about right, or not enough.
"Four weeks" is a starting assumption for someone with light SRE exposure, not a fact about the exam. Score yourself honestly against the six statements below — one point each, and only tick a box if it's true right now, unaided, not "I've read about this once."
| Score | Your lane | What to change about this plan |
|---|---|---|
| 5–6 | Lane A — two weeks | Halve every day count in the table below (round up). Skip the tool-by-tool tour in Module 5 and go straight to the four functional categories. Sit your first mock exam at the end of your Week 1, not Week 3. |
| 3–4 | Lane B — four weeks | Run the plan exactly as written below — this is who it's built for. |
| 0–2 | Lane C — six weeks | Read What is SRE?, SLIs, SLOs & error budgets and Toil & automation slowly first, in that order, before opening a single blueprint module. Then multiply every day count below by roughly 1.5. |
The eight modules — days, lessons, and blueprint pages
☺ Like you're 10: One row per quiz chapter — how many days to give it, which pages actually teach it, and which practice pile to drill afterward.
This is the plan's reference spine — everything after it is a calendar wrapped around these eight rows. The module order is fixed by DevOps Institute's own syllabus and by the blueprint pages' own numbering; read the linked lesson to build the idea properly, then use the blueprint page for exactly how DevOps Institute phrases and tests the same ground, then drill the matching practice bank cold.
If exam weighting were the axis, all eight rows would be identical, the same way The SREF Exam's own diagram deliberately draws eight equal spokes. Study load is a different axis: Module 2 pairs with three lesson pages because SLO math is the vocabulary every later module assumes you already have; Module 5 pairs with several tool pages because "sort a tool by function, not brand" only becomes reliable after you've actually looked at more than one tool. Modules 1, 3, 7 and 8 lean on one or two pages you likely already have partial context on. Don't read the shorter rows as less important — read them as already closer to done.
Weeks 1–2 — the first six modules
☺ Like you're 10: Week one covers the vocabulary everything else depends on. Week two covers watching the system and reacting to it.
Week 1 (Days 1–7) — orientation, Principles & Practices, SLOs & Error Budgets, and the start of Reducing Toil
Day 1: read The SREF Exam in full for format and registration context, skim the certifications overview, and run the pace check above if you haven't already. Days 2–3 (Module 1): What is SRE?, then the Module 1 blueprint page for exactly how DevOps Institute phrases the same ground. Days 4–6 (Module 2): SLIs, SLOs & error budgets first, slowly — this is the module every later one leans on — then SLO windows & composite SLOs, multi-window burn-rate alerting, and the Module 2 blueprint page. Close with a first, untimed pass at the SLOs & Toil practice bank. Day 7: start Module 3 — Toil & automation.
Pick any SLO you can find a real number for — your own service, a public status page, or an invented 99.9% target over a 30-day window. Work out the error budget in minutes of allowed downtime by hand, slowly, the way Sol would. Then do it again for 99.95% over the same window. The point isn't speed yet — it's that the arithmetic stops feeling unfamiliar before you ever see it on a timed paper.
Week 2 (Days 8–14) — Reducing Toil, Monitoring & SLIs, and Tools & Automation
Day 8: finish Module 3 — the Module 3 blueprint page and a first pass of the SLOs & Toil bank's toil questions specifically. Days 9–11 (Module 4): Monitoring & observability, then the Module 4 blueprint page and a re-read of multi-window burn-rate alerting from the signal-selection side rather than the SLO-math side. Days 12–14 (Module 5): the SRE toolchain overview, then one representative tool per functional category rather than all of them — Prometheus for metrics, Jaeger for tracing, PagerDuty for paging, Gremlin for chaos — plus the Module 5 blueprint page's automation-maturity ladder, which is the actual thing this module tests. Close the week with an untimed pass at the Monitoring & Tools practice bank.
The exam tests the category and the maturity rung, not the vendor. If you can correctly sort an unfamiliar tool's one-sentence description into "metrics," "tracing," "paging," or "chaos," and correctly place a described practice on the runbook-to-autonomous-remediation ladder, you don't need to have memorized all twenty-seven tool pages this course happens to carry.
Weeks 3–4 — the last two modules, then the exam itself
☺ Like you're 10: Week three finishes the reading. Week four stops reading altogether and starts proving it, five different times.
Week 3 (Days 15–21) — Anti-Fragility, Organizational Impact, and Frameworks & the Future
Days 15–17 (Module 6): Postmortems & blameless culture and Chaos engineering — read both, since the blueprint module fuses them — then chaos engineering at scale and the Module 6 blueprint page for the fragile/robust/antifragile distinction specifically. Days 18–19 (Module 7): Incident management & on-call, then SRE team topologies and the Module 7 blueprint page. Close with an untimed pass at the Resilience & Culture practice bank, covering both modules. Days 20–21 (Module 8): re-read What is SRE?'s DevOps/SRE section specifically, then reliability economics and the Module 8 blueprint page for how SRE sits against ITIL, Agile, and DORA.
You've now covered all eight modules at least once, with a practice-bank pass behind six of them. Week 4 is deliberately not about learning anything new — it's about finding out, under a real clock, which of the eight didn't stick as well as it felt like it did while reading.
Week 4 (Days 22–28) — consolidate, then five mock exams
| Day | What you do | Why |
|---|---|---|
| 22 | First full pass of the SREF Concept Reference and Know It Cold — SREF; skim Closed-Book Strategy — No Docs Map once. | Tighten the near-miss definitions a single-best-answer distractor is built to exploit, before a single mock sitting is on the record. |
| 23 | Mock Exam Set 1 — 60 minutes, closed-book, no notes. Score it by module, not just overall. | Set 1 is deliberately the recall baseline. A weak module here is a genuine content gap, not a pacing issue — start a short mistake log now. |
| 24 | Re-drill whichever module(s) scored weakest on Set 1, using the matching practice bank above; read Answer Triage — SREF for the general technique of eliminating options by what a term structurally can't be. | A day of repair while Set 1's specific misses are still fresh is worth more than moving straight to Set 2 with the same gap intact. |
| 25 | Mock Exam Set 2 + review. | A second, distinct 40 questions — later sets lean further from straightforward recall toward layered, multi-step scenarios. |
| 26 | Mock Exam Set 3 + review. | Same shape, further from recall — proves the vocabulary generalizes past the exact phrasing you first learned it in. |
| 27 | Mock Exam Set 4 + review. | By now a stable, repeated weak module (if any) should be obvious across three sittings, not just one. |
| 28 | Mock Exam Set 5 — the closest rehearsal to the real thing — then work through the readiness gates below. | The final dress rehearsal, sat with days still left to react to whatever it reveals. |
Five mock exams, spent without wasting them
☺ Like you're 10: A practice quiz for a closed-book test is like a bag of sweets — eat the same one twice and the second time you're just remembering the wrapper, not learning the flavor again.
Unlike the hands-on, performance-based exams elsewhere on this site — where doing the same category of task across several mocks is closer to genuinely fine, because you still have to produce the artifact from scratch each time — the SREF is a closed-book, single-best-answer paper. The moment you recognize the correct proposition on a repeated question — "the answer is B, that's the SLA one" — the paper stops measuring what you know and starts measuring what you remember about that specific sitting. This course's five sets exist precisely so you never have to re-sit the same numbered paper: each is a distinct 40 questions, evenly spread across the eight modules, scored against the same 65% pass mark, with later sets leaning progressively further from straightforward recall toward layered scenarios that combine two modules in one question stem.
Three things still genuinely waste a sitting even with five distinct papers available: sitting two sets back to back on the same day with no review between them, so neither miss gets fixed before the next paper repeats it; reading a worked answer explanation before you've genuinely committed to an answer yourself, which turns a mock into a copying exercise; and treating a strong overall score as proof of readiness when a single module quietly sat at 1 or 2 out of 5 the whole time. Score every sitting by module, not just as one overall percentage — a flat 75% can hide a module that's actually failing, and only the per-module breakdown tells you which one.
Keep it to one line per missed question, in a plain text file: module · what the question actually asked · which distractor tempted me · the one-sentence rule that would have stopped me. After Set 1 you'll have a handful of lines; by Set 5 you should be able to read back through them and find the same rule preventing the same mistake twice — that repetition, not a rising score alone, is the real signal that a gap has closed.
Readiness checkpoint — the gates before you book
☺ Like you're 10: "I feel ready" isn't a measurement — here are four things you can actually check before you spend money on a seat.
| # | Gate | How you check it | If it fails |
|---|---|---|---|
| 1 | Two of your last three sittings clear comfortably | Two of Sets 3–5 score 75% or better overall — ten points of margin over the 65% pass mark — and no single module drops to 2/5 or worse on either sitting. | Re-drill the practice bank for whichever module keeps dropping, then sit one more fresh set if you have one left. |
| 2 | All eight modules, named in order, from memory | Without looking: list all eight SREF blueprint modules in their syllabus order, in under a minute. | Read the "full map, in order" line near the top of any blueprint page a few more times — the sequence itself is free recall points on any question that references "the module after X." |
| 3 | The SLA/SLO/SLI and toil definitions come out unaided | No notes: define all three in one sentence each, in the fixed order they have to occur, and list three criteria for what counts as toil. | Drill Know It Cold — SREF and the Concept Reference until these stop requiring thought. |
| 4 | Your mistake log is quiet | Re-reading it, most rules should feel obvious now rather than surprising — a sign the gap actually closed rather than just aging out of recent memory. | Any rule that still surprises you belongs on one more focused re-drill before you book, not a general re-read of everything. |
"I've read every module" isn't on this list on purpose. Reading is input; the exam only scores output under a clock. If you've read all eight modules but never sat a single timed mock, you have prepared for a different exam than the one you're about to take.
Remy the Rabbit: Four weeks feels slow for a one-hour quiz. Can't I just read all eight pages this weekend and sit it Monday?
Professor Owl: You could read them. Reading isn't the same skill as recalling them cold, with a clock running and no tab to fall back on.
Sol the Sloth: And some of those pages carry more arithmetic than others. I'd rather give the error-budget module three unhurried days than rush it into one.
Timmy the Turtle: Which is exactly why the plan gives it three, and gives Frameworks & the Future two — not because one matters less on the exam, DevOps Institute never said that, but because one simply has more ground to cover first.
Remy the Rabbit: Fine. But I still get to sit five mock exams fast, right? Bang, bang, bang, all in one sitting?
Nutty the Squirrel: Space them out. Burn through all five back to back and you'll remember which one had the tricky postmortem question, not whether you actually know what makes a postmortem blameless.
Professor Owl: Four weeks, eight modules, five sittings with room to fix what each one reveals. Then the gates decide you're ready — not a feeling the night before.
That's the full plan — twenty module-study days sized to what each of the eight actually asks of you, a pace check for your own starting point, and a final week built around five distinct mock exams instead of one long re-read. Pair it with The SREF Exam for the day-of logistics, Closed-Book Strategy for exam-taking tactics specific to having no documentation at all, the Concept Reference and Know It Cold for the definitions a distractor is built to exploit, and the full practice bank whenever a single module needs more reps than its three focused sets provide. Then start at Module 1 if you're beginning fresh, or straight into whichever week above matches where your pace check actually placed you.
1. Why do the day counts in this plan differ across the eight modules, given that DevOps Institute doesn't publish per-module exam weighting? 2. You scored 4 out of 6 on the pace check — which lane are you in, and what do you do differently from the default plan? 3. Which two modules share the "SLOs & Toil" practice bank alongside SLOs & Error Budgets? 4. Why is repeating the same mock-exam paper within this four-week window a genuinely bad idea, when repeating the same category of hands-on lab elsewhere on this site is closer to fine? 5. Name two of the four readiness gates, and one gate this page deliberately leaves off the list.
Check your answers
- The days reflect how much course material each module pairs with — how much there is to read and drill — not a guess about how many exam questions it's worth. With no published weighting, this plan treats all eight modules as equally likely to appear and equally worth mastering; the varying day counts are about study load, not importance.
- Lane B — four weeks. Run the plan exactly as written below the pace check.
- Reducing Toil (Module 3) shares the SLOs & Toil practice bank with Module 2, alongside Module 1's own reference to the same bank.
- Because the SREF is closed-book and single-best-answer: recognizing a repeated question's correct proposition lets you answer it from memory of the paper itself rather than from actual knowledge, which stops measuring what you know. A hands-on lab still requires you to produce the artifact from scratch each time, so recognizing the task category doesn't let you skip the work.
- Any two of: two of your last three sittings clearing 75%+ with no module at 2/5 or worse; all eight modules named in order from memory; the SLA/SLO/SLI and toil definitions coming out unaided; a quiet mistake log. The gate this page deliberately leaves off is "I've read every module" — reading is input, and the exam only scores recall under a clock.