HomeRoadmaps › Kyverno Certified Associate
Active CNCF / Linux Foundation certification

Kyverno Certified Associate (KCA) Roadmap

A six-week, official-source path from Kubernetes-native policy fundamentals through Helm operations, CLI testing, resource selection, validate/mutate/generate and CEL authoring, signed-image controls, reports, exceptions, metrics, upgrades, and recovery.

90-minute online examMultiple choiceBeginnerValid 2 years50 weighted questions
Integrity and currency: Verified August 20, 2026 against official Linux Foundation, CNCF, public KCA curriculum, and current Kyverno documentation. The 50 questions are original practice—not copied, recalled, leaked, or live exam material. Product APIs evolve: the KCA curriculum names established ClusterPolicy concepts, while current Kyverno 1.19 documentation marks legacy policy types deprecated and recommends stable CEL-based policy types. Learn both the published objectives and the current migration direction.

Official blueprint translated into 50 questions

The official exam page lists six domains. This bank converts each percentage exactly into a 50-question allocation: 18% = 9, 18% = 9, 12% = 6, 10% = 5, 32% = 16, and 10% = 5.

Fundamentals · 18%
9 questions: policies, rules, YAML, admission, OCI
Install / configure / upgrade · 18%
9 questions: Helm, CRDs, flags, RBAC, HA, upgrades
Kyverno CLI · 12%
6 questions: apply, test, jp, CI context
Applying Policies · 10%
5 questions: cluster application, selection, settings
Writing Policies · 32%
16 questions: validation through CEL and cleanup
Policy Management · 10%
5 questions: reports, exceptions, metrics
1

Fundamentals, admission, YAML, and OCI identity

Week 1

Build a clean mental model before writing syntax. Separate Kubernetes authentication, authorization, mutation, validation, persistence, background reconciliation, and reporting.

  • Explain Kyverno as a declarative policy engine rather than a scheduler, runtime, CNI, or API-server replacement
  • Read policy resources as apiVersion, kind, metadata, spec, rules, selection, context, preconditions, and behavior
  • Distinguish namespaced and cluster scope and select the narrowest sufficient policy type
  • Trace an AdmissionReview through authentication, authorization, mutation, validation, persistence, and report intermediates
  • Explain why ordinary reads do not run through admission and why background scans lack the original actor context
  • Identify OCI registry, repository, tag, manifest, digest, signature, and attestation as separate concepts
  • Compare the public curriculum's legacy vocabulary with current stable CEL-based Kyverno policy types
  • Complete questions 1-9 and explain every distractor
2

Helm installation, CRDs, controllers, flags, RBAC, HA, and upgrades

Week 2

Operate Kyverno as a critical admission dependency. Know which controller owns each behavior and which ServiceAccount needs each permission.

  • Install a pinned official Helm chart in a dedicated Namespace and inventory all installed resources
  • Map the admission, background, reports, and cleanup controllers to webhooks, UpdateRequests, reports, and deletions
  • Inspect CRDs and distinguish policy resources from supporting and internal resources
  • Review ConfigMap keys, container flags, dynamically managed webhooks, resource filters, and Namespace selectors
  • Extend controller RBAC with explicit aggregated ClusterRoles instead of editing chart-owned roles
  • Understand concurrent admission replicas versus leader-elected reports and background processing
  • Model fail-closed security against availability and test protected system-Namespace recovery
  • Read every intervening release note, render the chart diff, back up policies and values, test CRD/API migration, and prepare rollback
  • Complete questions 10-18 and project 1
3

CLI apply, test, jp, and policy CI

Week 3

Make policy behavior repeatable before it reaches an admission webhook. Pin versions and model external context instead of relying on hidden cluster state.

  • Use kyverno apply against resource files and explore detailed results, reports, values, exceptions, and cluster mode
  • Use kyverno test with expected pass, fail, skip, mutation, generation, and exception outcomes
  • Use kyverno jp function, parse, and query to develop JMESPath safely
  • Supply values files, context, user information, parameter resources, target resources, and registry access only as required
  • Produce machine-readable CI output and fail when tests are missing or expected outcomes change
  • Test malformed input, absent optional fields, array variation, controller templates, exceptions, and dependency failure
  • Validate in a disposable cluster because CLI success does not prove webhook, RBAC, registry, TLS, or controller health
  • Complete questions 19-24 and build the project 2 test suite
4

Apply policies with deliberate scope and rollout

Week 4

Selection errors are policy errors. Learn when a request never reaches Kyverno, when the engine filters it, when a rule skips, and when a violation is audited or enforced.

  • Combine kinds, operations, names, Namespaces, selectors, namespaceSelectors, subjects, roles, match, and exclude carefully
  • Differentiate API-server webhook selectors from Kyverno engine resource filters and policy match logic
  • Use per-rule failureAction, policy failurePolicy, timeout, background, warnings, and reporting settings deliberately
  • Start broad new validation in Audit, inspect current-state reports, fix resources, approve narrow exceptions, and stage Enforce
  • Test selected, excluded, inapplicable, and request-only scenarios rather than only compliant resources
  • Protect labels used for scoping or break-glass from unauthorized mutation
  • Measure admission latency and errors before increasing scope
  • Complete questions 25-30
5

Write validate, mutate, generate, image, variable, autogen, cleanup, and CEL policies

Week 5

This is the largest official domain. Practice complete behavior chains, not isolated YAML fragments.

  • Write validation with patterns, anyPattern, anchors, deny conditions, foreach, Pod Security, and CEL
  • Use preconditions to skip rule bodies and mark request-dependent rules background false
  • Write idempotent strategic-merge mutations and precise RFC 6902 patches with correct JSON Pointer escaping
  • Understand mutation ordering, cascading rules, foreach element and elementIndex, and immutable Kubernetes fields
  • Generate resources from data, clone, or cloneList; decide synchronize and generateExisting lifecycle behavior
  • Grant the background controller exact permissions and inspect failed UpdateRequests
  • Verify image digests, trusted attestors, signatures, attestations, and registry credentials using safe lab images
  • Use AdmissionReview variables, ConfigMaps, Kubernetes API calls, service calls, registry context, and JMESPath with bounded data
  • Inspect autogen rules and cover suppression cases such as mixed kinds, selectors, and Pod JSON Patch mutation
  • Use current DeletingPolicy for new scheduled cleanup work while recognizing legacy cleanup and TTL concepts
  • Complete questions 31-45 and projects 2-3
6

Reports, exceptions, metrics, final review, and exam readiness

Week 6

Treat management evidence as a live operating system. A report result needs context, an exception needs governance, and a metric needs a service objective.

  • Interpret PolicyReport and ClusterPolicyReport results: pass, fail, warn, error, and skip
  • Remember that reports represent current resources and are not immutable blocked-request history
  • Enable exceptions only in a controlled Namespace and protect creation with RBAC and review
  • Scope exceptions to exact policy, target, condition, owner, rationale, compensating controls, and expiry
  • Monitor admission volume and latency, policy results and duration, controller reconciliation/requeues/drops, cleanup, API clients, and version
  • Rehearse troubleshooting across readiness, endpoints, certificates, RBAC, UpdateRequests, reports, metrics, events, logs, and release notes
  • Take a timed 50-question session inside 90 minutes, review every wrong answer, then repeat weak domains
  • Complete questions 46-50, review all 40 flashcards, and finalize project evidence

Three portfolio projects

1. Production-shaped installation

Helm, CRDs, controllers, flags, aggregated RBAC, HA behavior, metrics, fail-closed recovery, staged upgrade, and teardown.

Open project surface
2. Namespace policy pack

Selection, validate, preconditions, background, mutate, generate, JSON Patch, autogen, CEL, deletion, reports, exceptions, and CI.

Open project surface
3. Signed-image enforcement

Safe OCI images, digest identity, signature and attestation policy, CLI tests, admission, cache, failures, metrics, upgrade recovery, and trust revocation.

Open project surface

All PrepKloud learning surfaces

Official references

Linux Foundation KCA

Exam format, experience level, duration, validity, domains, competencies, registration, and resources.

Open official exam page
CNCF KCA

Program overview, official weights, and curriculum link.

Open CNCF page
CNCF curriculum

Public curriculum repository containing the KCA blueprint.

Open curriculum
Candidate Handbook

Linux Foundation exam policies and PSI proctoring requirements.

Open handbook
Kyverno introduction

Policy management purpose and core capabilities.

Open introduction
Installation and HA

Helm, components, availability, and security-versus-operability.

Open installation · HA
Configuration and upgrades

Flags, RBAC, filters, webhooks, certificates, release changes, and migration.

Configuration · Upgrades
Kyverno CLI

Official apply, test, and jp command references.

Open CLI reference
Validation and mutation

Patterns, deny, background, CEL, strategic merge, JSON Patch, and foreach.

Validate · Mutate
Generation and image verification

Data, clone, synchronization, UpdateRequests, signatures, attestations, and digests.

Generate · Verify images
Autogen, variables, and cleanup

Controller translation, context and API data, plus current deletion direction.

Autogen · External data
Reports, exceptions, and metrics

Policy outcomes, controlled bypasses, and operational telemetry.

Reports · Exceptions · Metrics

Frequently asked questions

Is the KCA active in 2026?

Yes. As verified August 20, 2026, CNCF and Linux Foundation publish the Kyverno Certified Associate as an active online proctored multiple-choice certification.

How long is the exam and how long is certification valid?

The official Linux Foundation KCA page lists a 90-minute exam, beginner experience level, and two-year certification validity. It also states a 12-month exam eligibility window and one retake for the current offering; verify purchase terms when registering.

Which domain deserves the most study time?

Writing Policies is 32%, the largest domain. It covers validation, preconditions, background scans, mutation, generation, image verification, variables and API calls, JSON patches, autogen, cleanup, and CEL. Do not neglect the two 18% domains.

Are the 50 questions real exam items?

No. They are original educational questions created from the public curriculum and official documentation. They contain no recalled, copied, leaked, or live exam material and cannot predict the exact exam.

Why does the roadmap discuss both ClusterPolicy and CEL policies?

The public KCA objectives name established ClusterPolicy concepts. Current Kyverno 1.19 documentation says CEL-based policy types have full feature parity and marks legacy policy types deprecated. Candidates need curriculum literacy and current operational judgment.

Can the projects be completed without a cloud account?

Yes. Use a disposable local cluster, local OCI registry, synthetic workloads, and non-production signing material. Optional hosted registries or clusters can add cost and data-handling concerns; verify current provider documentation before using them.

Independence and exam-integrity disclaimer: PrepKloud is independent and is not affiliated with or endorsed by the Linux Foundation, CNCF, or the Kyverno project. KCA, Kyverno, Kubernetes, and related marks belong to their respective owners. Exam objectives, versions, APIs, pricing, retake terms, policies, and documentation may change. Verify official sources before scheduling or operating production systems. The practice content is original and does not reproduce confidential or live exam questions. Passing a practice quiz, finishing a lab, or earning a certification does not prove production readiness, compliance, security, or employment outcomes.

Start with the weighted question bank

Take 50 questions under a 90-minute timer, review all explanations, drill 40 cards, then prove the concepts through three safe projects.