If you’ve seen “CNAPP experience preferred” in a cloud security job posting and quietly closed the tab, reopen it. A CNAPP (Cloud-Native Application Protection Platform) is a single platform that checks cloud configurations, workloads, identities, and infrastructure code together, then tells you which combination of problems an attacker could actually chain. You can learn CNAPP concepts without an enterprise license, and they’re less foreign than the acronyms suggest: almost every finding a CNAPP raises lands on something an administrator or engineer configures every day.
This guide covers what CNAPP is, what each acronym inside it does, how one famous breach maps onto it, and how to get hands-on for free.
What is CNAPP, without the vendor slides
Gartner coined the term in August 2021, in its research note Innovation Insight for Cloud-Native Application Protection Platforms, to describe a problem every cloud team already had: too many security tools that didn’t talk to each other. One tool checked account settings. Another scanned containers. A third looked at permissions. Each produced its own alert queue, and nobody saw the full picture.
A CNAPP consolidates those into one platform. The four core components are the acronyms you’ll hear in interviews:
| Component | What it does | The everyday task it resembles |
|---|---|---|
| CSPM — Cloud Security Posture Management | Continuously scans cloud accounts for misconfigurations and compliance drift | Config audits, baseline checks, “who opened this port?” |
| CWPP — Cloud Workload Protection Platform | Protects VMs, containers, and serverless functions, from build through runtime | Patching, hardening, endpoint agents |
| CIEM — Cloud Infrastructure Entitlement Management | Maps who and what can access which resources, and flags permissions nobody uses | Access reviews, AD group cleanup |
| IaC scanning — Infrastructure as Code scanning | Checks Terraform, CloudFormation, and similar templates before deployment | Change review before the change hits prod |
Most platforms now add two more layers: CDR (Cloud Detection and Response) for spotting suspicious activity as it happens, and AI-SPM (AI Security Posture Management) for the AI models and data pipelines companies now run in the cloud.
Look at the right-hand column again. None of it is exotic. That’s the point of this whole article.
Why companies buy one: context beats volume
The selling point isn’t more alerts. It’s fewer, better ones.
Picture separate tools reporting on the same environment:
- Tool A: Server X has a critical vulnerability.
- Tool B: Role Y has broad permissions.
- Tool C: Bucket Z contains customer data.
Each alert is true. Each one alone looks like something for next sprint. A CNAPP connects them into a graph and asks a different question: can these be chained? If Server X faces the internet, has an exploitable flaw, and runs with Role Y, and Role Y can read Bucket Z, those three “medium” findings become one critical attack path. Practitioners call this a toxic combination.
That’s the whole value proposition: network exposure, workload weakness, identity permission, and data sensitivity, evaluated together instead of in four separate queues.
If that sounds like threat modeling, it is. We broke down the manual version in how threat modeling works in practice. A CNAPP’s attack path analysis is a machine doing a narrow slice of that work, continuously. (And if “exploitable” versus “vulnerable” still feels fuzzy, the difference between a vulnerability and an exploit is exactly what attack path analysis depends on.)
The breach that explains CNAPP better than any datasheet: Capital One, 2019
The CNAPP category didn’t exist in 2019, so treat this as a counterfactual, not a product pitch. But the Capital One breach is the clearest example of the problem CNAPPs were designed to solve, because no single link in the chain was exotic.
What happened. In March 2019, an attacker who had previously worked as a systems engineer at AWS accessed Capital One’s AWS environment and copied data tied to roughly 100 million people in the US and about 6 million in Canada. The chain had three links:
- An exposed workload with a flaw. A web application firewall running on an EC2 instance was misconfigured in a way that allowed server-side request forgery (SSRF): the attacker could make the server send requests on her behalf.
- An identity with too much reach. She used that SSRF to query the instance metadata service and retrieve temporary credentials for the IAM role attached to the instance. That role could list and read S3 buckets the firewall had no operational reason to touch.
- Data within reach of that identity. With those credentials, she listed and copied the contents of hundreds of S3 folders.
Capital One didn’t catch it through its own monitoring. An outside party saw the attacker posting about the data on GitHub and reported it in July 2019, months after the access. Capital One later paid an $80 million civil penalty to the Office of the Comptroller of the Currency.
Where a CNAPP would have changed the odds:
- Attack path analysis would have flagged the combination before anyone exploited it: an internet-facing instance with a role that could read sensitive storage. A textbook toxic combination.
- CIEM would have shown the role’s S3 permissions going unused by the firewall’s normal behavior. That’s a right-sizing recommendation waiting to happen.
- CSPM would flag instances still accepting the older, token-less version of the metadata service. AWS released IMDSv2 in November 2019, after the breach; it requires a session token that blocks the simple SSRF used here. It still isn’t enforced unless someone sets it to required.
- Runtime detection could have raised an alert when a firewall’s credentials suddenly started enumerating storage buckets, behavior nothing in that server’s job description explains.
“Could have” is doing honest work in that last bullet. Detection depends on tuning, and tuning depends on people who know what normal looks like in their environment.
Look at what the chain was actually made of: a reverse proxy config, an instance role, a storage permission. Anyone who has administered servers has touched all three. The breach didn’t take genius-level skill to execute or to prevent. It was three ordinary configuration decisions that nobody evaluated together.
CNAPP through the Threat & Control Method
At Blue Team Academy, we teach defensive security through the Threat & Control Method: Inventory → Threats → Controls → Scale. A CNAPP is, conceptually, a tool that automates parts of each step.
- Inventory. A CNAPP’s first job is discovery: every account, workload, identity, and data store it can reach. First deployments routinely surface assets nobody had listed. That’s an Asset Inventory, generated continuously.
- Threats. Attack path analysis is a partial Threat Model. It computes which technical chains exist. What it can’t know is business context: which data matters most, which system can’t go down, which “exposed” server is a deliberate public endpoint.
- Controls. Right-sizing a role, enforcing IMDSv2, blocking a Terraform change that would open a bucket. Turning findings into a prioritized Security Plan is still a human decision.
- Scale. Policy-as-code and multi-cloud coverage let you apply the same reasoning across AWS, Azure, and Google Cloud: a Repeatable Decision Process instead of one-off fixes.
That’s why the tool alone doesn’t make anyone a cloud security engineer. The platform produces findings. The professional decides which ones matter, why, and in what order. Cybersecurity is not rocket science, but it does require knowing which question the dashboard is actually answering.
The platforms you’ll hear about (and the design choices behind them)
A handful of names come up again and again in cloud security conversations. As of November 2026:
- Wiz — known for an agentless approach: it connects through cloud provider APIs and scans workload snapshots instead of requiring software on every instance. Google completed its $32 billion acquisition of Wiz in March 2026; Wiz kept its brand and its support for AWS, Azure, and other clouds.
- Palo Alto Networks Cortex Cloud — the evolution of Prisma Cloud, announced in February 2025, which folds CNAPP capabilities into Palo Alto’s Cortex security operations platform alongside cloud detection and response.
- CrowdStrike Falcon Cloud Security — comes from an endpoint and runtime-protection background, using lightweight sensors for deep workload visibility.
- Microsoft Defender for Cloud — Microsoft’s native option, common in organizations already invested in Azure and Entra ID, with coverage extending to AWS and Google Cloud.
Don’t memorize feature lists. They change every quarter. Understand the design decisions behind them, because that’s what evaluation teams actually argue about.
How to evaluate a CNAPP like a practitioner
If you move into a cloud security role, helping evaluate or tune one of these platforms is a realistic early assignment. These five questions separate a useful evaluation from a polished demo:
- Agentless, agent-based, or both? API-based scanning gives fast, low-friction visibility. Agents (often eBPF-based sensors on Linux) give deeper runtime detail and the ability to block. Most mature platforms now offer both. The real question is where your environment needs which.
- Does it treat every cloud you run equally? Ask for the demo against the cloud that’s second most important to you, not the first.
- Does it fit the pipeline, or fight it? IaC and image scanning should plug into GitHub or GitLab workflows without making developers wait. A control that slows every deploy gets bypassed.
- How does it reduce noise? Don’t ask how many findings it generates. Ask how it decides what not to show you.
- Who owns the fix? A finding that lands in a queue nobody owns is just a well-formatted log entry. Ask how the platform routes remediation to the team that actually controls the resource.
Question five is where operational experience pays off. Anyone who has worked a ticket queue knows detection without ownership changes nothing.
How to learn CNAPP concepts for free
Enterprise CNAPP licenses aren’t priced for home labs. The underlying concepts, though, are all available through open-source tools and free practice environments.
Before you start: use a dedicated personal cloud account, never an employer’s, and set a billing alert on day one. Free tiers have limits, and a forgotten instance is how a free lab becomes a surprise invoice.
Step 1: Rebuild each CNAPP component with open-source tools
- CSPM → Prowler. Run it against your lab AWS, Azure, or Google Cloud account and read every finding. Prowler maps its checks to compliance frameworks, so you learn what auditors look for at the same time.
- IaC and image scanning → Checkov or Trivy. Point them at a Terraform file and a container image. Then deliberately write a bad Terraform resource, like a public bucket or a wide-open security group, and watch the scanner catch it.
- Runtime → Falco. Created by Sysdig and a CNCF graduated project since February 2024, Falco is the open-source reference point for Linux and Kubernetes runtime detection. Install it on a small cluster and trigger a rule by opening a shell inside a running container.
If containers are new territory, read our breakdown of container security beyond image scanning first.
Step 2: See the attacker’s side
- flAWS and flAWS2 are free, browser-based challenges that walk through common AWS misconfigurations. flAWS2 includes a defender track, which is the one to prioritize.
- CloudGoat (by Rhino Security Labs) deploys deliberately vulnerable AWS scenarios. Treat it like live ammunition: isolated lab account only, and tear it down when you’re done.
For a defender’s view of how attackers abuse legitimate cloud tooling, see how cloud administration commands get abused.
Step 3: Touch a commercial interface
Microsoft Defender for Cloud includes free foundational posture management for Azure subscriptions, which lets you explore a commercial CNAPP-style dashboard against your own lab environment.
A certification note, answered honestly: the AZ-500 exam that many older guides recommend retired on August 31, 2026, and Microsoft has named SC-500 as its replacement. If a specific certification is a hard requirement for the roles you’re targeting, check the current exam list on Microsoft Learn before paying for anything. If it isn’t, the lab work above builds the skill either way.
Step 4: Turn the lab into evidence
This is where most people stop too early. A finished lab isn’t a portfolio piece until it’s documented:
- Deploy a deliberately misconfigured environment with Terraform.
- Run Prowler and Checkov. Save the “before” reports.
- Fix the findings in code, not by clicking around the console.
- Re-run the scans. Save the “after” reports.
- Write up which findings you fixed first and why. That’s the part a hiring manager actually reads.
We cover the write-up side in how to document cybersecurity labs, and how to phrase the result on paper in our cybersecurity resume tips. A line like “Remediated IAM and storage misconfigurations in a lab AWS environment through Terraform, cutting high-severity posture findings to zero” reads very differently from “Familiar with cloud security tools.”
What the market data says (and what it doesn’t)
The U.S. Bureau of Labor Statistics projects employment of information security analysts to grow 21% from 2025 to 2035, much faster than the 3% average for all occupations, with a May 2025 median annual wage of $129,180 (BLS Occupational Outlook Handbook, last modified August 2026). BLS doesn’t track “cloud security engineer” or “CNAPP specialist” as separate occupations, so read those numbers as the broader security market, not a cloud-specific promise.
What the data can’t tell you is how long any individual transition takes. That depends on your starting role, your local market, and what you can show. CNAPP knowledge is one strong signal for cloud-facing security roles. It isn’t a shortcut around the rest.
The takeaway
CNAPP looks intimidating because of the acronyms. Underneath, it’s posture checks, workload hardening, access reviews, and change control, evaluated together instead of in separate silos. If you’ve run infrastructure, you already understand the parts. What you’re building is the habit of seeing them as one attack surface.
That’s blue team thinking in miniature, and it’s the same reasoning that runs through any realistic path into a security career. If you’re coming from an IT role and want the full Threat & Control Method with structured practice instead of piecing it together alone, the Blue Team Academy program walks through it step by step.
Prefer to start smaller? The Keep IT Safe newsletter breaks down real cloud incidents and defensive concepts like this one. Subscribe to Keep IT Safe.

