SRE Practitioner
The SRE Practitioner is the natural next stop after SRE Foundation (SREF) — this course's own anchor certification — and it comes from the same body, the DevOps Institute, now part of PeopleCert. It keeps the parts of the format that matter most for exam-day nerves identical: forty multiple-choice questions, 65% to pass. What changes is the sitting length, stretched from sixty minutes to ninety, and — more importantly — what the questions actually ask of you. SREF wants you to define an error budget correctly. Practitioner wants you to read a paragraph describing a specific organization's incident, its tooling, and its team shape, and decide what an SRE would actually do about it. Below: what it is, exactly what's published about format and cost, how it differs from SREF in substance and not just duration, and who should — and shouldn't — spend the money.
Imagine two tests about crossing the road safely. The first hands you a rulebook and asks you to recite it back: what does a red light mean, how long should you wait at a crosswalk, what's the stopping distance at 30 miles an hour. You pass by remembering facts. The second test doesn't ask you to recite anything — it drops you at a real, messy intersection with a broken signal, a school two blocks away, and a bus route that just changed, and asks what you'd actually do about it, in that order, with that specific mess. Same rulebook underneath both tests. The second one just makes you use it on a situation instead of reciting it from memory.
What it is, and how it differs from SRE Foundation in kind, not just length
☺ Like you're 10: Same publisher, same multiple-choice format, same passing bar — but this one hands you a situation instead of a definition.
SRE Practitioner is issued by the DevOps Institute, the same vendor-neutral body behind SRE Foundation, and it sits directly above it in the DevOps Institute's own certification ladder. It is still a knowledge-based, multiple-choice exam — there is no live terminal, no cluster to fix, nothing performance-graded the way this platform's Kubernetes-flavored certifications are. That distinction matters enough to say plainly: Practitioner is not a hands-on credential, and taking it will not teach you to debug a stuck rollout or read an etcd snapshot. What it changes relative to SREF is the shape of the questions within that same multiple-choice format. SREF is built to test recall and correct definition — can you identify what an SLI is, distinguish toil from ordinary operational work, name the roles in an incident response. Practitioner is built around applied scenarios: a described organization, its current maturity, its incident history, and its tooling gaps, with the question asking which SRE practice or model fits that specific context best. The syllabus reflects the shift directly — module titles like SRE Anti-patterns, Building Secure and Reliable Systems, and SRE & Incident Response Management are framed around diagnosing and choosing, not defining.
Practitioner tests which SRE model fits this organization, not what SRE means. If SREF is the vocabulary test, Practitioner is closer to a case-study exam — still closed to a keyboard, still multiple choice, but every question arrives wrapped in a scenario rather than posed as a bare definition.
Format, cost, and prerequisites — verify before you book
☺ Like you're 10: Forty questions, ninety minutes, 65% to pass, and a recommended — not required — badge to already have.
The table below reflects what the DevOps Institute and PeopleCert publish for SRE Practitioner as of this writing. Treat every number here as a snapshot: this exam family has changed publisher, pricing currency, and validity period within the last few years, and it will likely change again.
| Item | SRE Practitioner |
|---|---|
| Issuing body | DevOps Institute, now part of PeopleCert |
| Exam code | None officially published — the name is service-marked, not code-numbered the way DOFD is for DevOps Foundation. Any SREFND- or DOICSRE-style code you're quoted is a training partner's internal course code, not an exam code. |
| Format | Multiple choice, web-based, online-proctored via PeopleCert |
| Questions | 40 — identical to SRE Foundation |
| Duration | 90 minutes — up from SREF's 60 |
| Passing score | 65% — identical to SRE Foundation. (Some third-party exam-dump sites quote 70%; the official pass mark is 65%.) |
| Price | Exam voucher priced from roughly £216 on PeopleCert's own GBP pricing page; your currency, local tax, and any bundled training will differ |
| Validity | PeopleCert's standard 3-year CPD cycle — older articles and course brochures written before the PeopleCert acquisition still quote a 2-year cycle; that's out of date |
| Prerequisites | PeopleCert states no formal prerequisite is enforced at booking. In practice, most accredited training partners present SRE Foundation as a strong recommendation before Practitioner — treat it as expected, not mandatory, and confirm what your specific training provider requires |
| Languages | Historically a narrower set than SREF — English, Japanese, and Portuguese (Brazil) have been offered; confirm current language availability before you register |
Two things worth flagging beyond the usual "prices change" caveat. First, sources disagree on whether this exam is open-book or closed-book — this course's own certifications page describes SRE Foundation as closed-book, while other PeopleCert-sourced material describes the same exam family as open-book, online-proctored. Confirm the current policy directly on the official exam page before you sit; walking in prepared for the wrong policy is an easy, avoidable way to lose time on the day. Second, the DevOps Institute's own published SRE Practitioner Exam Study Guide is, to a striking degree, a curated reading list of material that's free elsewhere — Google's SRE Book and Workbook, PagerDuty's freely published incident-response documentation, and the Principles of Chaos Engineering. If you've worked carefully through the equivalent pages on this course, you are closer to exam-ready than the price tag might suggest. Confirm everything — price, format, prerequisites, validity — on the official DevOps Institute SRE Practitioner page before you pay.
Same shape, different depth: SREF and Practitioner side by side
☺ Like you're 10: Picture two boxes joined by a dotted "recommended" arrow — same test format in each box, but the right-hand one asks harder, messier questions in the extra thirty minutes it's given.
Laid out next to each other, the two exams look almost identical on paper and diverge entirely in what they're actually testing:
| SRE Foundation (SREF) | SRE Practitioner | |
|---|---|---|
| Question | Can you define the vocabulary correctly? | Can you apply it to a specific organization's mess? |
| Format | 40 questions, multiple choice | 40 questions, multiple choice |
| Duration | 60 minutes | 90 minutes |
| Pass mark | 65% | 65% |
| Syllabus shape | 8 modules, definitional — SLIs/SLOs, toil, monitoring, tools, anti-fragility, org impact, frameworks | 8 modules, applied — anti-patterns, secure & reliable systems, full-stack observability, tooling & automation, incident response, chaos engineering |
| Prerequisite | None | None enforced; SREF strongly recommended by most training partners |
| Where this course covers it | The 8-module SREF blueprint | This page, plus the deep-dive and reference pages linked below |
What the eight modules actually ask of you
☺ Like you're 10: Eight topics again, like SREF — but this time each one is "given this mess, what do you do," not "what does this word mean."
DevOps Institute doesn't publish a per-module weighting for Practitioner any more than it does for SREF — there's no legitimate "module 4 is worth 15% of the exam" figure to study toward, whatever a third-party prep site tells you. What is published is the list of eight modules, and the pattern across them is consistent: each one is framed as a diagnosis or a decision rather than a definition.
| # | Module (as published) | What it's really asking |
|---|---|---|
| 1 | SRE Anti-patterns | Spot the org that's copying SRE's vocabulary without its mechanisms — the "renamed ops team" failure mode this course's own What is SRE? warns about |
| 2 | SLO is the Proxy for Customer Happiness | Given a described product and its users, pick the SLI that actually tracks whether a human is happy, not just whether a server is up |
| 3 | Building Secure and Reliable Systems | Reason about where security and reliability practice overlap and where they trade off against each other inside one system |
| 4 | Full Stack Observability | Choose the right signal — metric, log, or trace — for a described failure mode, not just recite the three pillars |
| 5 | SRE Tools and Automation, including AIOps | Match an operational problem to the right category of tooling, and reason about where automated triage helps versus where it hides a real signal |
| 6 | SRE & Incident Response Management | Given a described incident, assign the right role — Incident Commander, Scribe, Subject Matter Expert — and the right next action |
| 7 | Chaos Engineering | Decide whether a described chaos experiment is safe to run, and what blast-radius controls it's missing |
| 8 | SRE is the Purest Form of DevOps | Place SRE correctly relative to DevOps and ITIL when a scenario blurs the boundary between them |
To make the shift concrete: a SREF-style question might ask you to pick the correct definition of an error budget from four options. A Practitioner-style question is more likely to read like this — "A payments team's SLO is being met on paper, but the on-call engineer reports being paged eleven times last week for the same recurring alert with no user-facing impact. What should the team do first?" — and expect you to recognize that as an alert-fatigue and toil problem before anything else, the same diagnostic instinct this course builds across toil & automation and incident management & on-call. Nobody hands you that scenario word-for-word on the real exam — the DevOps Institute doesn't publish sample items — but that's the register the syllabus is written in, and it's worth rehearsing that style of reasoning, not just the definitions, before you sit it.
How this maps onto this course
☺ Like you're 10: You don't need a new textbook — this course already covers most of what Practitioner tests, just filed under different page titles.
Because this course is built around the same underlying discipline SREF and Practitioner both test, most of the eight modules land squarely on pages you can already read here, some at the foundational level and some at the depth Practitioner actually expects.
| Practitioner territory | Read on this course |
|---|---|
| Anti-patterns & renamed-ops-team failure modes | What is SRE? · SRE Anti-Patterns & Pitfalls |
| SLOs as a proxy for customer happiness | SLIs, SLOs & error budgets · SREF Module 2 blueprint · SLO windows & composite SLOs |
| Security and reliability overlap | Security's Overlap with Reliability |
| Full-stack observability | Monitoring & observability · Multi-window, multi-burn-rate alerting · Alert design & alert fatigue |
| Tools & automation | Toil & automation · The SRE toolchain |
| Incident response management | Incident management & on-call · Incident Command for Large-Scale Incidents · Case study: an outage post-mortem |
| Chaos engineering | Chaos engineering · Chaos Engineering at Scale · Gremlin · Gremlin Enterprise Chaos Engineering Certification for the hands-on counterpart |
| SRE, DevOps & frameworks | What is SRE? (the DevOps-vs-SRE section) · SREF Module 8 blueprint |
| Organizational models & team topologies | Organizational Impact of SRE · SRE Team Topologies |
Notice what's missing from that list: nothing here teaches you to sit a proctored exam or reproduces DevOps Institute's own scenario items, which the vendor doesn't publish. What this course gives you instead is the underlying judgment the scenarios are testing — read the pages above the way you'd study for Practitioner, not the way you'd cram a vocabulary list for SREF.
Before you register for anything, try writing one Practitioner-style question yourself about a service you actually operate. Pick a real recurring page or a real recent incident, then write: (1) one sentence describing the organizational context — team size, how mature the on-call rotation is, what's already automated; (2) three plausible next actions, only one of which is actually correct given that context; (3) the one-sentence reason the other two are traps. If you can write a convincing wrong answer as well as the right one, you understand the material at Practitioner's depth — reciting the right answer alone only proves SREF-level recall.
Who should sit it, and who should skip it
☺ Like you're 10: Good for someone accountable for reliability across several services who already knows the vocabulary; skippable if you're still learning the words, or if what you actually need is proof you can fix a real system.
Sit it if you're already accountable for reliability across more than one service — a senior SRE, an engineering manager who owns an on-call rotation, or a platform lead who needs to make the case for SRE mechanisms to a leadership team that doesn't yet have the vocabulary. It's also a reasonable next step if you've already sat SREF and want a credential that goes slightly deeper without committing to a hands-on, lab-based exam. And it's worth having on a CV specifically because it's the DevOps Institute's most advanced SRE-branded credential — for enterprise procurement and some hiring filters, that seniority signal is the whole point, independent of how much new knowledge the exam itself imparts.
Skip it, at least for now, if any of these fit. You haven't sat SRE Foundation yet — sit that first; Practitioner assumes the vocabulary rather than re-teaching it, and most training partners will tell you the same. You want proof you can actually operate a system under pressure rather than reason about one on paper — a hands-on chaos-engineering credential like Gremlin's Enterprise Chaos Engineering Certification, or a performance-based exam like the CNCF's CKA, will teach and prove a different, more concrete skill. And if your actual gap is a specific cloud platform's operational tooling rather than SRE practice in the abstract, the certifications hub covers Google Cloud's and AWS's more operationally-flavored exams, which will serve you better than a second knowledge-based DevOps Institute badge.
Foxy: I already passed SREF last quarter. Is Practitioner just SREF again with a longer clock?
Professor Owl: No — same forty questions, same 65% bar, but the questions stop asking "what is a blameless postmortem" and start describing an org with a specific incident history and asking what you'd actually do about it.
Timmy the Turtle: Before either of you books anything — has anyone here confirmed the current price and the open-book policy on the vendor's own page? This site's own pages don't even agree with each other on the book-policy question.
Benny the Beaver: I automated that check once. It broke the week PeopleCert changed their pricing page structure. Now I just look it up by hand, every time, right before I book.
Foxy: Fine. And the scenario style — is that actually harder, or just longer?
Professor Owl: Harder in a specific way. Definitions have one right answer. A scenario question usually has three plausible answers and one correct one — and the exam is testing whether you can tell them apart under a described organization's real constraints, not in the abstract.
1. What stays identical between SRE Foundation and SRE Practitioner, and what changes? 2. Is SRE Practitioner a hands-on, performance-based exam like the CKA, or a knowledge-based one? 3. Is SRE Foundation a strict prerequisite for SRE Practitioner? 4. Name two of the eight Practitioner modules and describe, in your own words, the kind of question each is likely to ask. 5. Name one alternative certification worth considering instead of Practitioner, and the specific gap it fills that Practitioner doesn't.
Check your answers
- Identical: 40 questions, multiple choice, 65% to pass. Changed: the duration extends from 60 to 90 minutes, and the questions shift from defining terms to applying SRE practice to a described organizational scenario.
- Knowledge-based. It's still a multiple-choice, web-based, proctored exam with no live terminal or graded cluster — the "scenario-heavy" framing describes the style of the questions, not the format of the assessment.
- No — DevOps Institute enforces no formal prerequisite at booking. In practice, most accredited training partners strongly recommend SREF first, since Practitioner assumes its vocabulary rather than re-teaching it.
- Any two of: SRE Anti-patterns (spot an org copying SRE's vocabulary without its mechanisms); SLO is the Proxy for Customer Happiness (pick the SLI that tracks real user happiness for a described product); Building Secure and Reliable Systems (reason about security/reliability trade-offs in one system); Full Stack Observability (choose the right signal for a described failure); SRE Tools and Automation/AIOps (match a problem to the right tooling category); SRE & Incident Response Management (assign the right role and next action in a described incident); Chaos Engineering (judge whether a described experiment is safe to run); SRE is the Purest Form of DevOps (place SRE correctly against DevOps/ITIL in a blurred scenario).
- Any one of: Gremlin's Enterprise Chaos Engineering Certification, for hands-on proof of chaos-engineering skill rather than reasoning about it on paper; the CKA, for hands-on Kubernetes operations skill; Google Cloud's Professional Cloud DevOps Engineer or AWS's CloudOps Engineer Associate (both on the certifications hub), for reliability practice tied to a specific cloud platform's tooling rather than tested in the abstract.