Certifications for site reliability engineering
There is no single "SRE certification" the way there's a single CKA for Kubernetes administration — the discipline is young enough, and cuts across enough clouds and practices, that the useful credentials come from three different places testing three different things. One is a cloud vendor's professional exam that happens to be organized around SRE practice. One is a vendor-neutral knowledge test about the practice itself, with no cloud attached. One is a cloud operations exam that touches reliability without naming SRE at all. None of them is a checklist to collect in order — each is a different tool, and the right one depends on whether you're proving breadth, depth, or a specific cloud's operational muscle memory.
Imagine three different badges you could earn for "knows how to keep a treehouse running." One badge is from the company that built your treehouse's specific pulley system — it proves you know that pulley system well, including a chapter on "how much wobble is too much wobble." Another badge is from an independent judge who never sold you any pulleys — it only asks whether you understand the idea of budgeting for wobble at all, regardless of whose treehouse. The third badge is about running the treehouse day to day on one particular company's cloud-based treehouse kit — patching leaks, watching the weather, filing the maintenance log. All three are real. None of them, alone, proves you can keep a treehouse standing forever.
The landscape
☺ Like you're 10: One badge is cloud-specific and deep; one is vendor-neutral and about the ideas only; one is cloud-specific and about day-to-day operations. Pick based on what you're trying to prove.
Two axes are worth separating before you look at any single exam. The first is vendor-neutral versus vendor-specific: the SRE Foundation tests the practice itself — SLOs, error budgets, toil, blameless postmortems — independent of any cloud, while the Google Cloud and AWS credentials test that same territory (or adjacent territory) filtered through one vendor's specific services and console. The second axis is knowledge-based versus scenario-driven: all three of these happen to be multiple-choice exams with no live terminal, but the Google Cloud exam leans on scenario-driven case studies that ask you to apply a practice to a described situation, not just define it.
For most people new to the discipline, the SRE Foundation is the natural first step — it's the shortest, the cheapest, and the only one of the three that teaches the vocabulary this whole course is built around without also requiring you to learn a cloud console. The two cloud-specific exams are then a question of which cloud your platform actually runs on, not which one is "better."
Google Cloud Professional Cloud DevOps Engineer
☺ Like you're 10: Google invented the term "SRE," and this is their grown-up badge for building the delivery pipeline and keeping the result reliable, using Google's own tools.
The Professional Cloud DevOps Engineer is issued by Google Cloud. Because Google's own engineers wrote the original Site Reliability Engineering book and effectively coined the term, this is the closest thing the market has to an "official" SRE-flavored certification — one of its five domains is literally titled "applying site reliability engineering practices to a service." It sits in Google's Professional tier, above the Associate and foundational credentials, and is aimed at practitioners who both build CI/CD pipelines and operate what those pipelines ship.
It's for engineers who work in Google Cloud day to day — the exam tests Cloud Build, Cloud Deploy, Cloud Monitoring and Cloud Logging by name, not portable concepts alone, so someone who has never touched GCP will spend most of their study time learning a console rather than reinforcing SRE judgment they already have.
| Item | Detail |
|---|---|
| Format | Multiple-choice and multiple-select, including scenario-driven case studies |
| Question count | Not published by Google — verify at the vendor's own page |
| Duration | 2 hours |
| Passing score | Not published by Google — verify at the vendor's own page |
| Prerequisites | None formally required; Google recommends 1+ years of hands-on Google Cloud experience |
| Domain | Weight |
|---|---|
| Bootstrapping a Google Cloud organization | 20% |
| Building and implementing CI/CD pipelines | 25% |
| Applying site reliability engineering practices to a service | 18% |
| Implementing service monitoring strategies (observability and troubleshooting) | 25% |
| Optimizing performance and cost | 12% |
The overlap with this course is direct: the SRE-practices domain maps onto measuring reliability, the observability domain onto monitoring and observability, and the CI/CD domain onto release engineering and progressive delivery — read those three before you open Google's exam guide, and the guide will read like vocabulary you already have, filtered through Google's specific product names.
SRE Foundation (SREF)
☺ Like you're 10: No cloud, no console — just a quiz on the ideas: SLOs, error budgets, toil, and blameless postmortems.
The SRE Foundation is issued by the DevOps Institute. It's a closed-book, knowledge-based, vendor-neutral certification specifically about SRE as a practice — SLOs, error budgets, toil, and blameless postmortems — rather than any one cloud platform's implementation of those ideas. Where the Google Cloud exam spends most of its weight on pipelines and a specific console, the SRE Foundation spends all of its weight on the practice itself, which makes it the exam most directly aligned with what this course teaches.
It's aimed at anyone starting out in the discipline, or anyone who wants a portable, cloud-agnostic proof of SRE literacy that doesn't expire in relevance the moment they change employers or clouds. A follow-on SRE Practitioner certification exists from the same body for those who want to go deeper after the Foundation.
| Item | Detail |
|---|---|
| Format | Multiple-choice, closed-book |
| Question count | 40 |
| Duration | 60 minutes |
| Passing score | 65% |
| Prerequisites | None formally required |
This exam's territory is essentially this course's early curriculum: SLIs, SLOs & error budgets, toil and automation, and postmortems and blameless culture cover the SRE Foundation's core vocabulary directly — someone who's read those three pages carefully is close to exam-ready without opening a single cloud console.
↗ DevOps Institute official page
AWS Certified CloudOps Engineer – Associate (SOA-C03)
☺ Like you're 10: AWS's badge for running production day to day — monitoring, fixing, and keeping things reliable on AWS specifically. It used to have a different name; the old name is retired now.
The AWS Certified CloudOps Engineer – Associate, exam code SOA-C03, is issued by AWS. It doesn't call itself an "SRE" exam, but its domains — monitoring and remediation, reliability and business continuity, deployment automation — are squarely reliability-operations territory for anyone whose platform runs on AWS. It's aimed at practitioners who operate and troubleshoot AWS workloads day to day, closer in spirit to a systems-administration exam than a software-engineering one.
Naming note, and it matters: this exam replaced the older "AWS Certified SysOps Administrator – Associate" exam, which was retired — its last bookable day was 29 September 2025. Any study material that still refers to "SysOps Administrator Associate" as the current name is describing an exam you can no longer sit; treat it as out of date rather than as an alternate name for the same thing.
| Item | Detail |
|---|---|
| Format | Multiple-choice and multiple-response |
| Question count | 65 total — 50 scored, 15 unscored |
| Duration | 130 minutes |
| Passing score | Scaled score of 720 out of 1000 |
| Prerequisites | None formally required |
| Domain | Weight |
|---|---|
| Monitoring, Logging and Remediation | 20% |
| Reliability and Business Continuity | 16% |
| Deployment, Provisioning and Automation | 18% |
| Security and Compliance | 16% |
| Networking and Content Delivery | 18% |
| Cost and Performance Optimization | 12% |
The clearest overlap with this course sits in three places: monitoring and observability for the largest domain, capacity planning and performance for the reliability and cost domains, and reliability patterns for the business-continuity domain's emphasis on redundancy and failover design.
Which one should you take?
☺ Like you're 10: New to SRE — start with the vendor-neutral one. Already working on a cloud — take that cloud's one. Want both breadth and depth — do the neutral one first either way.
| You are... | Best next step | Why |
|---|---|---|
| New to SRE, no cloud preference yet | SRE Foundation | Cheapest, shortest, and it's the only one of the three that teaches the practice with nothing else attached — the ideas transfer to whichever cloud you end up on. |
| Already operating production on Google Cloud | Professional Cloud DevOps Engineer | Structured tour of your own environment, with the SRE domain formalizing vocabulary you're likely already using loosely. |
| Already operating production on AWS | CloudOps Engineer Associate (SOA-C03) | Direct coverage of the monitoring, remediation and business-continuity work that's already your day job. |
| Want to prove both the practice and a cloud | SRE Foundation first, then the matching cloud exam | The Foundation gives you the vocabulary the cloud exam assumes; taking the cloud exam first means learning the concept and the console names at the same time, which is slower. |
| Not on Google Cloud or AWS at all | SRE Foundation | The only one of the three with no cloud dependency — everything you learn stays useful if your employer or your cloud changes. |
Exam codes, prices, question counts and passing scores move without much announcement, and vendors retire and rename exams — the SOA-C03 naming note above is exactly that kind of change, caught mid-transition. Everything on this page was sourced from each vendor's own certification page as of August 2026. Verify current logistics at the vendor's own page before booking — this page is a snapshot, not a live feed.
Foxy: Okay, I want to actually prove I know SRE, not just that I know one cloud's version of it. Which one do I take?
Professor Owl: Start with the SRE Foundation. Forty questions, an hour, closed-book, and it's entirely about the practice — SLOs, error budgets, toil, blameless postmortems. No cloud console in sight.
Timmy: And then check which cloud you actually run production on before you spend more money. If it's Google Cloud, the Professional Cloud DevOps Engineer has a whole domain on SRE practice worth 18% — but half the exam is Cloud Build and Cloud Monitoring specifics.
Foxy: We're on AWS.
Timmy: Then it's the CloudOps Engineer Associate — SOA-C03. And don't study from anything that calls it "SysOps Administrator Associate." That name stopped being bookable on 29 September 2025.
Professor Owl: None of the three replace the others. They prove different things to different readers — pick based on what you need to prove, not which badge looks most impressive on a profile.
1. Which of the three certifications is vendor-neutral, and what does that mean for how it stays relevant if you change employers? 2. What's the current name of the AWS exam covered here, and what name should you treat as out of date? 3. Which certification has a domain explicitly titled around applying SRE practice to a service, and what's that domain's weight? 4. If you're new to SRE with no cloud preference yet, which certification should you take first, and why?
Check your answers
- The SRE Foundation (SREF), from the DevOps Institute. Because it tests the practice itself rather than any cloud's implementation, what you learn for it stays useful even if you change clouds or employers — nothing about it is tied to a specific vendor's console.
- The current name is AWS Certified CloudOps Engineer – Associate (SOA-C03). Treat "AWS Certified SysOps Administrator – Associate" as out of date — that exam was retired, with its last bookable day on 29 September 2025.
- The Google Cloud Professional Cloud DevOps Engineer, whose "Applying site reliability engineering practices to a service" domain carries 18% of the exam weight.
- The SRE Foundation. It's the cheapest and shortest of the three, and it's the only one that teaches the SRE vocabulary and practice without also requiring you to learn a specific cloud vendor's console and service names.