ExamCert

ok hear me out.

everyone talks about DP-600 like it's this sweeping "Fabric platform" credential. Microsoft calls it the Analytics Engineer Associate cert. the marketing makes it sound like you'll walk out knowing lakehouses, pipelines, real-time intelligence, the whole ecosystem.

you won't. you'll walk out knowing DAX. and semantic models. and Power BI, wearing a Fabric-branded trench coat.

look at the actual domain breakdown:

Plan, implement, and manage a solution for data analytics — 10-15%

Prepare and serve data — 40-45%

Implement and manage semantic models — 20-25%

Explore and analyze data — 20-25%

do the math. "Prepare and serve data" plus "semantic models" plus "explore and analyze" is basically 85-90% of the exam. and that whole cluster is Power BI's world — DAX measures, star schemas, row-level security on semantic models, report performance tuning. the part that's genuinely "new Fabric platform knowledge" (workspace setup, capacity planning, governance) is stuck at 10-15%.

so when a Fabric skeptic on reddit says "DP600 has a good component of Power BI (DAX in particular)," that's underselling it. it's not a component. it's the spine of the exam.

the format, for reference: 40-60 questions, 100-120 minutes, 700/1000 to pass, $165 USD. one guy on reddit failed with 672 and passed the very next day with an 850 — which tells you this exam rewards knowing exactly where you're weak, not grinding everything equally.

why this actually matters for you

if you're a Power BI developer who's never touched a lakehouse, you are MUCH closer to passing this than you think. the DAX and modeling knowledge you already have covers most of the exam surface. don't let the "Fabric" branding scare you into thinking you need six months of lakehouse architecture study first.

and if you're a data engineer who's great with Spark and pipelines but weak on DAX — you are in more trouble than you think. i've seen people assume DP-600 is the "engineering" fabric cert and DP-700 is the "analyst" one. it's the opposite. DP-700 is the engineering-heavy one. DP-600 will absolutely eat you alive on semantic model design if your DAX is rusty.

the trap nobody warns you about

"Prepare and serve data" sounds generic but it's where most failures happen, because it blends dataflows, T-SQL views in the warehouse, AND semantic model prep into one bucket. you need to know when to push transformation logic upstream into a dataflow versus downstream into a DAX measure versus into a SQL view — and the exam expects you to pick the right layer for the right scenario, not just get a correct number.

star schema design questions show up constantly too. if you can't look at a flat table and instantly see where the fact/dimension split should happen, budget extra time here.

is it worth doing before or after DP-700?

most people who've done both say clear DP-600 first if you're coming from a BI background, because the shared Fabric fundamentals (workspaces, OneLake, licensing) carry over and DP-700 gets noticeably easier once you've already internalized the platform basics. if you're coming from a hardcore engineering background instead, honestly either order works — just don't assume DP-600 is the "lighter" one. it isn't lighter, it's just differently hard. the DP-600 exam overview has the full current domain weightings if you want to double check against whatever guide you're using right now, since Microsoft does update these percentages.

study reality check

figure 6-10 weeks if you're doing this alongside a full-time job, less if you already live in Power BI daily. the official Microsoft learning path is a fine baseline but it will NOT prepare you for how specific the scenario questions get about which Fabric item does what. drilling scenario-style questions matters more than reading docs here — that's basically true for every Fabric exam honestly.

i ran a bunch of practice rounds through ExamCert to stress-test my semantic modeling knowledge specifically, since that's the domain that actually decides pass/fail for most people. $4.99 for lifetime access, money-back if it doesn't help — cheap enough that it's not a real decision, just do it. if you want to test the waters first there's also a free DP-600 practice test you can run through before committing to anything.

Hot take to open with: the KCNA is the most *misunderstood* certification in the entire Kubernetes lineup.

Half the internet says it's a waste of money. The other half sells you a 40-hour course for it. Both are wrong, and I have opinions. Seven of them.

1. It's multiple choice. It is NOT a lab exam.

This trips up so many people. The CKA, CKAD and CKS drop you into a real terminal and make you do things. The KCNA does not. It's 60 multiple-choice questions in 90 minutes, online proctored, and you need 75% to pass.

So if you've been grinding `kubectl` muscle memory for the KCNA — that's lovely, genuinely useful, and mostly irrelevant to this specific exam. The KCNA wants to know if you understand *concepts*. What a Service does. Why you'd want a ReplicaSet. What the CNCF landscape even is. Which means a set of free KCNA practice questions is a far better use of your first evening than a cluster.

2. 75% is higher than people expect.

Everyone hears "associate level" and mentally files it next to a 65% pass mark. Nope. Three quarters. Out of 60 questions you can miss 15. That's not a lot of runway for "eh, I'll wing the ecosystem questions."

3. One domain is nearly half the exam and people still spread their time evenly.

The published curriculum weights look roughly like this:

Kubernetes Fundamentals — 44%

Container Orchestration — 28%

Cloud Native Application Delivery — 16%

Cloud Native Architecture — 12%

(Observability got folded in rather than living as its own bucket — double-check the current curriculum page before you book, the Linux Foundation moves these around.)

Fundamentals and Orchestration are 72% of the exam together. If you know Pods, Deployments, Services, ConfigMaps, Secrets, namespaces, the API server / scheduler / controller-manager / etcd control plane split, kubelet, and container runtime basics cold — you're most of the way to a pass before you've even opened the CNCF landscape diagram.

I'd honestly just start with a timed set of free KCNA practice questions on day one, see which of those two big domains is weaker, and dump your time there. Studying in weight order is boring advice but it's the advice that works.

4. The ecosystem questions are vibes-based and you can't cram them.

There's a chunk of this exam that's essentially "which CNCF project does the thing." Prometheus for metrics. Envoy for proxying. Helm for packaging. OpenTelemetry for traces. Argo for GitOps-flavoured delivery. Fluentd for logs.

You cannot reason your way to these. You either recognise the name or you don't. Make a stupid little flashcard deck of ~30 CNCF project names and what they do, review it on the bus for two weeks, done. That's the whole strategy. It feels like trivia because it kind of is.

5. The exam includes a free retake, which changes how you should think about booking.

$250 USD, and it comes with one free retake. So the actual risk of booking "a bit early" is lower than people treat it. Book it. Having a date on the calendar is worth about 30% more study effort than "I'll take it when I'm ready," which is a date that does not exist.

The certification is good for three years, for what it's worth.

6. It's a genuinely good on-ramp — but only if you're actually on-ramping.

Here's my slightly unpopular opinion: if you already run production Kubernetes, skip the KCNA. Go straight to the CKA or CKAD. You will be bored and $250 poorer.

But if you're a dev who keeps nodding along in infra meetings, or an ops person moving off VMs, or you're job hunting and need something on the CV that says "yes I understand containers" — it's great. It's the cheapest, fastest way to make cloud native stop feeling like a wall of jargon. It also forces you to learn the *why* before the CKA drowns you in the *how*.

7. Practice tests are where the actual learning happens, and you don't need to spend $200 on them.

For a concept exam, drilling questions IS the study method. Every question you get wrong teaches you a distinction you didn't know existed — like the difference between a Deployment and a StatefulSet, or when a DaemonSet is the right answer, or why a Service and an Ingress are not the same thing no matter how many blog posts blur them.

I used ExamCert for this because the pricing is faintly ridiculous — $4.99 for lifetime access, with a 100% money-back guarantee if you don't pass. Compared to the $200–$300 practice bundles floating around, it's not a close call. Spend the saved money on the exam voucher.

---

Realistic timeline if you're starting from roughly zero: three to four weeks, an hour a day. Two weeks on Fundamentals and Orchestration, one week on delivery and architecture plus your CNCF flashcards, and a final week of nothing but timed practice runs.

Then book it, pass it, and go get the CKA — which is where the terminal actually bites back.

I have been using ExamCert practice exams to prep for cloud and security certifications. Sharing the resources that helped me most.

All exams and free practice tests are on the homepage: https://www.examcert.app/

The AWS Certified Cloud Practitioner practice questions were especially helpful: https://www.examcert.app/exams/aws-clf-c02/

The SC-900, Microsoft's Security, Compliance, and Identity Fundamentals exam, is a fantastic entry point into the security side of the Microsoft cloud. It is beginner-friendly, but "fundamentals" does not mean "no studying." Here are seven things I genuinely wish someone had told me before I sat down. If you want to sanity-check your readiness, the SC-900 exam page on ExamCert is a solid place to start.

1. It's conceptual, not hands-on

This is not a configuration exam. You will not be asked to click through a portal or write policy. SC-900 tests whether you can *describe* concepts and capabilities: what a service does, when you would use it, and how the pieces fit together. If you come from a hands-on background, resist the urge to over-engineer your answers. The exam wants the textbook definition, not your clever real-world workaround.

2. "Azure AD" is now "Microsoft Entra ID"

Microsoft rebranded Azure Active Directory to Microsoft Entra ID, and the whole identity family now lives under the Entra umbrella. Study materials written a while ago still say "Azure AD," and the concepts are identical, but the current exam uses Entra naming. Learn the new terms: Microsoft Entra ID, Entra ID Governance, Entra Permissions Management, and so on. Getting tripped up by a name you technically already know would be a frustrating way to lose points.

3. Identity is the biggest and most important domain

The four knowledge areas are roughly: core security/compliance/identity concepts, Microsoft Entra identity capabilities, Microsoft security solutions, and Microsoft compliance solutions. The identity domain carries serious weight, so do not skimp on it. Know the difference between authentication and authorization cold. Understand single sign-on, multifactor authentication, Conditional Access, passwordless options, and the four pillars of identity governance. If you only deeply learn one domain, make it this one.

4. Know the "defense in depth" and "Zero Trust" models

The conceptual domain leans hard on frameworks. Zero Trust and its guiding principles (verify explicitly, use least-privilege access, assume breach) come up repeatedly, and you should be able to recognize them in a scenario. Same with defense in depth and its layers, plus the shared responsibility model, which defines who secures what across IaaS, PaaS, and SaaS. These are easy points if you memorize the models and genuinely painful if you wing it.

5. So. Many. Product names.

The security and compliance domains are essentially a tour of Microsoft's portfolio. Microsoft Defender for Cloud, Microsoft Sentinel, Microsoft Defender XDR, Microsoft Purview (compliance, data lifecycle management, information protection, insider risk, eDiscovery), Microsoft Priva, and more. The trick is not memorizing every feature but building a one-line mental "what problem does this solve?" summary for each. When a question describes a need, you match it to the product. Practicing with question sets is the fastest way to burn these mappings in, and working through free SC-900 practice questions exposes you to the exact phrasing the exam uses.

6. Purview vs. Defender is the classic mix-up

Here is the distinction that saves you: Defender products protect against threats (attacks, malware, suspicious sign-ins), while Purview products govern and protect data and compliance (classification, retention, sensitivity labels, eDiscovery, insider risk). If a scenario is about stopping a bad actor, think Defender. If it is about labeling sensitive data, meeting a regulation, or investigating internal risk, think Purview. That single line resolved a huge share of the questions I found tricky.

7. It's genuinely passable in a week or two

SC-900 is one of the more approachable Microsoft certifications, and you do not need months. A focused one-to-two-week plan works for most people: read the official learning paths, watch a walkthrough or two, and then drill practice questions until you consistently score comfortably above the passing bar. The exam is multiple choice with no hands-on labs, so once the concepts and product names click, you are ready. Do not overthink it or let the long product list scare you into endless studying.

Bonus: treat it as a launchpad

SC-900 is a fundamentals cert, which means its real value is as a foundation. It pairs naturally with the identity-focused SC-300, the security operations SC-200, and the security engineering SC-100 down the line. Passing it gives you the shared vocabulary those tougher exams assume you already have. So even if a fundamentals badge is not your end goal, the concepts you lock in here pay off for every Microsoft security certification that follows.