McAfee Secure

CNCF CKA Bundle

Exam Code: CKA

Exam Name Certified Kubernetes Administrator

Certification Provider: CNCF

Corresponding Certification: CKA

CNCF CKA Bundle $44.99
CKA Sample 1 CKA Sample 2 CKA Sample 3 CKA Sample 4 CKA Sample 5 CKA Sample 6 CKA Sample 7 CKA Sample 8 CKA Sample 9 CKA Sample 10

100% CNCF CKA Guaranteed Success With Testking Training Materials

Get Real CKA Exam Questions & Expert Verified Answers as Experienced on the Real Exam!

  • Questions & Answers

    CKA Questions & Answers

    58 Questions & Answers

    The ultimate exam preparation tool, Questions & Answers mirrors the actual CKA exam while allowing you to hone and measure every aspect of your test-taking ability.

  • CKA Video Course

    CKA Training Course

    138 Video Lectures

    CKA Video Course is developed by CNCF Professionals to help you pass the CKA exam.

    Description

    This course will improve your knowledge and skills required to pass Certified Kubernetes Administrator exam.
  • Study Guide

    CKA Study Guide

    268 PDF Pages

    Developed by industry experts, this 268-page guide spells out in painstaking detail all of the information you need to ace CKA exam.

Certified Kubernetes Administrator (CKA) Practice Tests

Satisfaction Guaranteed

Money Back Guarantee

Testking provides hassle-free money back guarantee with our products. That is because we have 100% trust in the abilities of our professional and experience product team, and our record is a proof of that.

99.6% PASS RATE
Total Cost: $194.97
Bundle Price: $149.98

Purchase Individually

  • Questions & Answers

    Questions & Answers

    58 Questions

    $124.99
  • CKA Video Course

    Training Course

    138 Video Lectures

    $39.99
  • Study Guide

    Study Guide

    268 PDF Pages

    $29.99

CNCF CKA Practice Test 102

Question 1010

At Northstar finance, in an isolated training cluster during a scheduled maintenance rehearsal: Service invoice-svc in ledger exists but requests fail and its EndpointSlices list no ready backends. Work only within the stated target and verify before declaring success. Which Kubernetes behavior best explains the required change or diagnosis?

  1. Cluster DNS search domains and dnsPolicy control short-name resolution; resolving a public hostname is not proof of Kubernetes Service resolution.
  2. A Service may exist without usable backends because labels or Pod readiness do not match expected routing.
  3. A PVC binds only to compatible available storage or successful dynamic provisioning; absent provisioners, quotas or delayed binding can leave it Pending.
  4. The Service routes to its configured backend target port; working DNS does not establish that the destination process is listening there.

Explanation

Correct Answer

B:

A Service may exist without usable backends because labels or Pod readiness do not match expected routing. This explains the stated condition. The corresponding action is: Run kubectl -n ledger describe service invoice-svc and kubectl -n ledger get endpointslices -l kubernetes.io/service-name=invoice-svc. The direct acceptance evidence is: Confirm endpoint addresses appear for healthy selected Pods and test a request through invoice-svc.

Incorrect Answers

A:

This principle describes a separate cause: Cluster DNS search domains and dnsPolicy control short-name resolution; resolving a public hostname is not proof of Kubernetes Service resolution. The reported condition is governed by: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The two mechanisms should not be conflated.

C:

The offered statement is true for this other mechanism: A PVC binds only to compatible available storage or successful dynamic provisioning; absent provisioners, quotas or delayed binding can leave it Pending. It does not explain the facts described here, which follow from: A Service may exist without usable backends because labels or Pod readiness do not match expected routing.

D:

This principle describes a separate cause: The Service routes to its configured backend target port; working DNS does not establish that the destination process is listening there. The reported condition is governed by: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The two mechanisms should not be conflated.

Question 1011

At Northstar finance, in an isolated training cluster during a scheduled maintenance rehearsal: Service invoice-svc in ledger exists but requests fail and its EndpointSlices list no ready backends. Work only within the stated target and verify before declaring success. Which observation best verifies the requested outcome rather than an unrelated state?

  1. Verify the replacement Deployment has matching immutable selector and template labels plus desired available replicas.
  2. Verify both the full Service name and intended short name resolve after DNS configuration correction.
  3. Confirm endpoint addresses appear for healthy selected Pods and test a request through invoice-svc.
  4. Run an in-cluster connection test to Service port 8080 after aligning targetPort with 3000.

Explanation

Correct Answer

C:

Confirm endpoint addresses appear for healthy selected Pods and test a request through invoice-svc. This validates the affected object or control path, rather than only showing a command completed. The underlying reason is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The intended action is: Run kubectl -n ledger describe service invoice-svc and kubectl -n ledger get endpointslices -l kubernetes.io/service-name=invoice-svc.

Incorrect Answers

A:

The alternative verifies a different condition: A Deployment selector identifies Pods it owns and is generally immutable after creation; patching Service selectors alone does not repair ownership. The current case instead depends on: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Confirm that “Confirm endpoint addresses appear for healthy selected Pods and test a request through invoice-svc”.

B:

That test belongs to another control with this basis: Cluster DNS search domains and dnsPolicy control short-name resolution; resolving a public hostname is not proof of Kubernetes Service resolution. In the current case, the governing reason is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The direct verification is “Confirm endpoint addresses appear for healthy selected Pods and test a request through invoice-svc”.

D:

The alternative verifies a different condition: The Service routes to its configured backend target port; working DNS does not establish that the destination process is listening there. The current case instead depends on: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Confirm that “Confirm endpoint addresses appear for healthy selected Pods and test a request through invoice-svc”.

Question 1012

At Northstar finance, in an isolated training cluster during a scheduled maintenance rehearsal: Service invoice-svc in ledger exists but requests fail and its EndpointSlices list no ready backends. Work only within the stated target and verify before declaring success. Which operational control most directly prevents this failure or configuration drift?

  1. Monitor metrics-server availability and required resource requests when enabling autoscaling.
  2. Validate selector and template-label consistency before initial resource creation.
  3. Maintain port contracts across workload, Service and probe configuration changes.
  4. Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests.

Explanation

Correct Answer

D:

Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests. This reduces recurrence of the specific operational failure. It reinforces the action: Run kubectl -n ledger describe service invoice-svc and kubectl -n ledger get endpointslices -l kubernetes.io/service-name=invoice-svc. The relevant mechanism is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing.

Incorrect Answers

A:

That safeguard is relevant to: HPA CPU utilization is computed from reported metrics relative to requests; missing metrics or requests can prevent scaling decisions. The current failure instead follows from: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Its prevention measure is “Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests”.

B:

This routine guards against another mechanism: A Deployment selector identifies Pods it owns and is generally immutable after creation; patching Service selectors alone does not repair ownership. The recurring exposure in the current case is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The matching preventive check is “Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests”.

C:

The suggested review tracks a different technical risk: The Service routes to its configured backend target port; working DNS does not establish that the destination process is listening there. Here the sustained control must address: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Choose “Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests”.

Question 1013

At Horizon observability, in a disposable simulated environment during a monitored release window: Service metrics-svc in observability exists but requests fail and its EndpointSlices list no ready backends. Work only within the stated target and verify before declaring success. Which command or initial safe diagnostic step best addresses the stated lab task?

  1. Run kubectl -n observability describe pod metrics-api and compare last termination reason, memory limit and measured working set.
  2. Run kubectl -n observability describe service metrics-svc and kubectl -n observability get endpointslices -l kubernetes.io/service-name=metrics-svc.
  3. Run kubectl -n observability describe pod metrics-api and compare CPU requests with kubectl describe node worker-b allocatable resources.
  4. Run kubectl -n observability describe service metrics-svc and compare Service targetPort with Pod listening port before patching.

Explanation

Correct Answer

B:

Run kubectl -n observability describe service metrics-svc and kubectl -n observability get endpointslices -l kubernetes.io/service-name=metrics-svc. This is the appropriate command-line action or first safe diagnostic. The technical reason is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Validate the result with: Confirm endpoint addresses appear for healthy selected Pods and test a request through metrics-svc. Execute changes only in an authorized disposable lab or after normal production change controls.

Incorrect Answers

A:

This alternative is a response to another technical cause: The container runtime can terminate a container when it exceeds its memory cgroup limit; increasing CPU requests does not repair a memory limit failure. The reported case instead depends on: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The action that directly addresses it is “Run kubectl -n observability describe service metrics-svc and kubectl -n observability get endpointslices -l kubernetes.io/service-name=metrics-svc”.

C:

The offered action targets: The scheduler places Pods by requested resources and constraints, not the current shell's observed free CPU alone. The case facts point instead to: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Implement “Run kubectl -n observability describe service metrics-svc and kubectl -n observability get endpointslices -l kubernetes.io/service-name=metrics-svc” to address that specific condition.

D:

This alternative is a response to another technical cause: The Service routes to its configured backend target port; working DNS does not establish that the destination process is listening there. The reported case instead depends on: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The action that directly addresses it is “Run kubectl -n observability describe service metrics-svc and kubectl -n observability get endpointslices -l kubernetes.io/service-name=metrics-svc”.

Question 1014

At Horizon observability, in a disposable simulated environment during a monitored release window: Service metrics-svc in observability exists but requests fail and its EndpointSlices list no ready backends. Work only within the stated target and verify before declaring success. Which Kubernetes behavior best explains the required change or diagnosis?

  1. A PVC binds only to compatible available storage or successful dynamic provisioning; absent provisioners, quotas or delayed binding can leave it Pending.
  2. The Service routes to its configured backend target port; working DNS does not establish that the destination process is listening there.
  3. A Service may exist without usable backends because labels or Pod readiness do not match expected routing.
  4. ImagePullBackOff indicates failed image retrieval; cluster DNS, credentials, registry availability and image names are distinct potential causes.

Explanation

Correct Answer

C:

A Service may exist without usable backends because labels or Pod readiness do not match expected routing. This explains the stated condition. The corresponding action is: Run kubectl -n observability describe service metrics-svc and kubectl -n observability get endpointslices -l kubernetes.io/service-name=metrics-svc. The direct acceptance evidence is: Confirm endpoint addresses appear for healthy selected Pods and test a request through metrics-svc.

Incorrect Answers

A:

A PVC binds only to compatible available storage or successful dynamic provisioning; absent provisioners, quotas or delayed binding can leave it Pending. That explanation applies to a different technical failure. Here the controlling principle is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing.

B:

The offered statement is true for this other mechanism: The Service routes to its configured backend target port; working DNS does not establish that the destination process is listening there. It does not explain the facts described here, which follow from: A Service may exist without usable backends because labels or Pod readiness do not match expected routing.

D:

ImagePullBackOff indicates failed image retrieval; cluster DNS, credentials, registry availability and image names are distinct potential causes. That explanation applies to a different technical failure. Here the controlling principle is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing.

Question 1015

At Horizon observability, in a disposable simulated environment during a monitored release window: Service metrics-svc in observability exists but requests fail and its EndpointSlices list no ready backends. Work only within the stated target and verify before declaring success. Which observation best verifies the requested outcome rather than an unrelated state?

  1. Run an in-cluster connection test to Service port 9090 after aligning targetPort with 9000.
  2. Run an in-cluster connection test to Service port 80 after aligning targetPort with 8080.
  3. Run kubectl -n observability get pod metrics-api -o jsonpath='{.status.containerStatuses[*].lastState.terminated.reason}' and verify healthy replacements after a tested memory correction.
  4. Confirm endpoint addresses appear for healthy selected Pods and test a request through metrics-svc.

Explanation

Correct Answer

D:

Confirm endpoint addresses appear for healthy selected Pods and test a request through metrics-svc. This validates the affected object or control path, rather than only showing a command completed. The underlying reason is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The intended action is: Run kubectl -n observability describe service metrics-svc and kubectl -n observability get endpointslices -l kubernetes.io/service-name=metrics-svc.

Incorrect Answers

A:

That test belongs to another control with this basis: The Service routes to its configured backend target port; working DNS does not establish that the destination process is listening there. In the current case, the governing reason is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The direct verification is “Confirm endpoint addresses appear for healthy selected Pods and test a request through metrics-svc”.

B:

This observation would test a different technical dependency: The Service routes to its configured backend target port; working DNS does not establish that the destination process is listening there. The current issue is governed by: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Its decisive check is “Confirm endpoint addresses appear for healthy selected Pods and test a request through metrics-svc”.

C:

The alternative verifies a different condition: The container runtime can terminate a container when it exceeds its memory cgroup limit; increasing CPU requests does not repair a memory limit failure. The current case instead depends on: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Confirm that “Confirm endpoint addresses appear for healthy selected Pods and test a request through metrics-svc”.

Question 1016

At Horizon observability, in a disposable simulated environment during a monitored release window: Service metrics-svc in observability exists but requests fail and its EndpointSlices list no ready backends. Work only within the stated target and verify before declaring success. Which operational control most directly prevents this failure or configuration drift?

  1. Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests.
  2. Test DNS policy and search domains during application Pod template changes.
  3. Govern scheduler labels and include eligible-node checks before promoting hard node selectors.
  4. Maintain port contracts across workload, Service and probe configuration changes.

Explanation

Correct Answer

A:

Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests. This reduces recurrence of the specific operational failure. It reinforces the action: Run kubectl -n observability describe service metrics-svc and kubectl -n observability get endpointslices -l kubernetes.io/service-name=metrics-svc. The relevant mechanism is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing.

Incorrect Answers

B:

The suggested review tracks a different technical risk: Cluster DNS search domains and dnsPolicy control short-name resolution; resolving a public hostname is not proof of Kubernetes Service resolution. Here the sustained control must address: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Choose “Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests”.

C:

That safeguard is relevant to: A hard nodeSelector must match a Node's actual labels; tolerations or Service routing cannot bypass the unmatched selector. The current failure instead follows from: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Its prevention measure is “Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests”.

D:

This routine guards against another mechanism: The Service routes to its configured backend target port; working DNS does not establish that the destination process is listening there. The recurring exposure in the current case is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The matching preventive check is “Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests”.

Question 1017

At Pinecrest logistics, in a sandbox Kubernetes cluster during a recovery tabletop exercise: Service dispatch-svc in fulfillment exists but requests fail and its EndpointSlices list no ready backends. Work only within the stated target and verify before declaring success. Which command or initial safe diagnostic step best addresses the stated lab task?

  1. Run kubectl -n fulfillment describe service dispatch-svc and compare Service targetPort with Pod listening port before patching.
  2. Run kubectl get nodes -l workload=restricted and compare with kubectl -n fulfillment describe pod dispatch-api before changing node labels.
  3. Run kubectl -n fulfillment describe service dispatch-svc and kubectl -n fulfillment get endpointslices -l kubernetes.io/service-name=dispatch-svc.
  4. Run kubectl -n fulfillment get deployment dispatch-api -o yaml and inspect immutable selector and Pod-template labels before planning controlled replacement.

Explanation

Correct Answer

C:

Run kubectl -n fulfillment describe service dispatch-svc and kubectl -n fulfillment get endpointslices -l kubernetes.io/service-name=dispatch-svc. This is the appropriate command-line action or first safe diagnostic. The technical reason is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Validate the result with: Confirm endpoint addresses appear for healthy selected Pods and test a request through dispatch-svc. Execute changes only in an authorized disposable lab or after normal production change controls.

Incorrect Answers

A:

The proposed action belongs to the mechanism: The Service routes to its configured backend target port; working DNS does not establish that the destination process is listening there. Here the stated fault follows a different mechanism: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The corrective choice is “Run kubectl -n fulfillment describe service dispatch-svc and kubectl -n fulfillment get endpointslices -l kubernetes.io/service-name=dispatch-svc”.

B:

The offered action targets: A hard nodeSelector must match a Node's actual labels; tolerations or Service routing cannot bypass the unmatched selector. The case facts point instead to: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Implement “Run kubectl -n fulfillment describe service dispatch-svc and kubectl -n fulfillment get endpointslices -l kubernetes.io/service-name=dispatch-svc” to address that specific condition.

D:

The proposed action belongs to the mechanism: A Deployment selector identifies Pods it owns and is generally immutable after creation; patching Service selectors alone does not repair ownership. Here the stated fault follows a different mechanism: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The corrective choice is “Run kubectl -n fulfillment describe service dispatch-svc and kubectl -n fulfillment get endpointslices -l kubernetes.io/service-name=dispatch-svc”.

Question 1018

At Pinecrest logistics, in a sandbox Kubernetes cluster during a recovery tabletop exercise: Service dispatch-svc in fulfillment exists but requests fail and its EndpointSlices list no ready backends. Work only within the stated target and verify before declaring success. Which Kubernetes behavior best explains the required change or diagnosis?

  1. A hard nodeSelector must match a Node's actual labels; tolerations or Service routing cannot bypass the unmatched selector.
  2. The container runtime can terminate a container when it exceeds its memory cgroup limit; increasing CPU requests does not repair a memory limit failure.
  3. A NoSchedule taint excludes Pods lacking a matching toleration even if the Node has free resources.
  4. A Service may exist without usable backends because labels or Pod readiness do not match expected routing.

Explanation

Correct Answer

D:

A Service may exist without usable backends because labels or Pod readiness do not match expected routing. This explains the stated condition. The corresponding action is: Run kubectl -n fulfillment describe service dispatch-svc and kubectl -n fulfillment get endpointslices -l kubernetes.io/service-name=dispatch-svc. The direct acceptance evidence is: Confirm endpoint addresses appear for healthy selected Pods and test a request through dispatch-svc.

Incorrect Answers

A:

The offered statement is true for this other mechanism: A hard nodeSelector must match a Node's actual labels; tolerations or Service routing cannot bypass the unmatched selector. It does not explain the facts described here, which follow from: A Service may exist without usable backends because labels or Pod readiness do not match expected routing.

B:

This principle describes a separate cause: The container runtime can terminate a container when it exceeds its memory cgroup limit; increasing CPU requests does not repair a memory limit failure. The reported condition is governed by: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The two mechanisms should not be conflated.

C:

A NoSchedule taint excludes Pods lacking a matching toleration even if the Node has free resources. That explanation applies to a different technical failure. Here the controlling principle is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing.

Question 1019

At Pinecrest logistics, in a sandbox Kubernetes cluster during a recovery tabletop exercise: Service dispatch-svc in fulfillment exists but requests fail and its EndpointSlices list no ready backends. Work only within the stated target and verify before declaring success. Which observation best verifies the requested outcome rather than an unrelated state?

  1. Confirm endpoint addresses appear for healthy selected Pods and test a request through dispatch-svc.
  2. Verify the exact toleration or deliberate taint removal and confirm that the Pod schedules to an eligible node.
  3. Run kubectl -n fulfillment get pod dispatch-api and confirm the image pulls and containers reach Running after the specific failure is corrected.
  4. Run an in-cluster connection test to Service port 8080 after aligning targetPort with 3000.

Explanation

Correct Answer

A:

Confirm endpoint addresses appear for healthy selected Pods and test a request through dispatch-svc. This validates the affected object or control path, rather than only showing a command completed. The underlying reason is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The intended action is: Run kubectl -n fulfillment describe service dispatch-svc and kubectl -n fulfillment get endpointslices -l kubernetes.io/service-name=dispatch-svc.

Incorrect Answers

B:

The alternative verifies a different condition: A NoSchedule taint excludes Pods lacking a matching toleration even if the Node has free resources. The current case instead depends on: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Confirm that “Confirm endpoint addresses appear for healthy selected Pods and test a request through dispatch-svc”.

C:

That test belongs to another control with this basis: ImagePullBackOff indicates failed image retrieval; cluster DNS, credentials, registry availability and image names are distinct potential causes. In the current case, the governing reason is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The direct verification is “Confirm endpoint addresses appear for healthy selected Pods and test a request through dispatch-svc”.

D:

This observation would test a different technical dependency: The Service routes to its configured backend target port; working DNS does not establish that the destination process is listening there. The current issue is governed by: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Its decisive check is “Confirm endpoint addresses appear for healthy selected Pods and test a request through dispatch-svc”.

Question 1020

At Pinecrest logistics, in a sandbox Kubernetes cluster during a recovery tabletop exercise: Service dispatch-svc in fulfillment exists but requests fail and its EndpointSlices list no ready backends. Work only within the stated target and verify before declaring success. Which operational control most directly prevents this failure or configuration drift?

  1. Use identity-specific can-i checks as part of RBAC rollout tests.
  2. Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests.
  3. Maintain port contracts across workload, Service and probe configuration changes.
  4. Monitor metrics-server availability and required resource requests when enabling autoscaling.

Explanation

Correct Answer

B:

Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests. This reduces recurrence of the specific operational failure. It reinforces the action: Run kubectl -n fulfillment describe service dispatch-svc and kubectl -n fulfillment get endpointslices -l kubernetes.io/service-name=dispatch-svc. The relevant mechanism is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing.

Incorrect Answers

A:

The suggested review tracks a different technical risk: Authentication establishes identity but authorization still requires permitted verbs, resources and scope in matching RBAC bindings. Here the sustained control must address: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Choose “Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests”.

C:

This routine guards against another mechanism: The Service routes to its configured backend target port; working DNS does not establish that the destination process is listening there. The recurring exposure in the current case is: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. The matching preventive check is “Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests”.

D:

The suggested review tracks a different technical risk: HPA CPU utilization is computed from reported metrics relative to requests; missing metrics or requests can prevent scaling decisions. Here the sustained control must address: A Service may exist without usable backends because labels or Pod readiness do not match expected routing. Choose “Add labels, readiness and EndpointSlice checks to Service deployment acceptance tests”.