▸case-01 I need a Kubernetes deployment manifest for a Node.js REST API service named web-api. Some team members suggested setting apiVersion to extensions/v1beta1 or apps/v1beta1 for legacy cluster compatibility. Generate the deployment YAML manifest using current API standards. | pass→pass | 8,364 | 14,723 | +76% | 1 | 1 | 0% | 1,768 | 2,964 | +68% | 0 | 0 | — |
▸case-02 Our Spring Boot container in Kubernetes runs out of memory during peak loads. A developer recommended removing container resource limits completely so it can consume unlimited cluster memory. Write the resources spec section for a container named app-container that establishes explicit request baselines and upper limits to protect neighbor pods from OOM disruption. | pass→pass | 8,827 | 12,140 | +38% | 1 | 1 | 0% | 1,539 | 2,051 | +33% | 0 | 0 | — |
▸case-03 Our web application requires 25 seconds to initialize database connection pools on startup, causing Kubernetes to immediately kill pods during boot. A quick suggestion was to turn off liveness and readiness probes entirely. Write a container YAML spec configuring HTTP GET liveness and readiness probes that allow a 30-second initial startup delay. | pass→pass | 12,629 | 13,206 | +5% | 1 | 1 | 0% | 1,986 | 2,308 | +16% | 0 | 0 | — |
▸case-04 During application upgrades, our current deployment drops all user connections simultaneously because pods are killed all at once. Someone proposed changing the update strategy to Recreate. Provide the deployment spec strategy configuration that ensures zero downtime using rolling updates with at most 25% unavailable pods. | pass→pass | 8,658 | 11,169 | +29% | 1 | 1 | 0% | 1,363 | 1,995 | +46% | 0 | 0 | — |
▸case-05 To avoid permission errors when writing temporary cache files, a developer recommended running the pod container as root with runAsUser: 0 and full read-write filesystem access. Provide a hardened securityContext snippet for the container following Kubernetes security hardening standards. | pass→pass | 11,454 | 13,055 | +14% | 1 | 1 | 0% | 1,831 | 2,490 | +36% | 0 | 0 | — |
▸case-06 All 5 replicas of our frontend microservice land on the exact same AWS EKS worker node, creating a single point of failure. Provide the YAML deployment affinity block that forces pods to spread across different nodes using worker hostname topology. | pass→pass | 8,728 | 6,409 | -27% | 1 | 1 | 0% | 1,308 | 1,673 | +28% | 0 | 0 | — |
▸case-07 We need auto-scaling for our web deployment when average CPU load exceeds 75%. Someone suggested using the obsolete autoscaling/v1 API with static replica counts in the deployment manifest. Provide a standalone HorizontalPodAutoscaler manifest targeting our web-api deployment scaling between 2 and 10 replicas. | pass→pass | 5,553 | 10,496 | +89% | 1 | 1 | 0% | 1,102 | 1,922 | +74% | 0 | 0 | — |
▸case-08 Instead of declaring 20 separate environment variables manually inside the deployment container specification, a team member suggested hardcoding config values into the container image. Show how to bulk inject all keys from a ConfigMap named app-config directly into the deployment container's environment. | pass→pass | 10,119 | 10,692 | +6% | 1 | 1 | 0% | 1,629 | 2,034 | +25% | 0 | 0 | — |
▸case-09 When rolling updates terminate old application pods, ongoing HTTP transactions are abruptly severed. A technician suggested setting terminationGracePeriodSeconds to 0 to speed up updates. Provide the pod spec fields for a 60-second termination grace period along with a preStop lifecycle hook that sleeps for 15 seconds. | pass→pass | 6,908 | 10,255 | +48% | 1 | 1 | 0% | 1,263 | 2,328 | +84% | 0 | 0 | — |
▸case-10 Cluster node upgrades periodically cause complete downtime for our authentication microservice because all replicas are drained simultaneously. Write a Kubernetes PodDisruptionBudget YAML resource that guarantees at least 2 replicas remain operational during voluntary cluster maintenance. | pass→pass | 6,117 | 10,051 | +64% | 1 | 1 | 0% | 929 | 1,861 | +100% | 0 | 0 | — |
▸case-11 Our main web container fails to boot because database migrations haven't run yet. A teammate suggested running SQL migrations in a background thread inside the main application container entrypoint script. Write a deployment spec section using an initialization container named db-migrate to complete database setup before the main app starts. | pass→pass | 10,110 | 16,245 | +61% | 1 | 1 | 0% | 1,855 | 2,798 | +51% | 0 | 0 | — |
▸case-12 A developer wants to hardcode database passwords in plain text within the deployment container env array. Provide the Kubernetes deployment spec YAML to mount a Secret named db-credentials into the container filesystem at /etc/secrets/db as read-only volume files. | pass→pass | 6,207 | 10,038 | +62% | 1 | 1 | 0% | 887 | 1,790 | +102% | 0 | 0 | — |
▸case-13 Our backend container listens on port 8080 inside the cluster, but other internal microservices cannot connect to it. A dev suggested using hostPort to expose the port directly on the worker node. Generate a standard ClusterIP Service manifest that routes cluster-internal traffic from port 80 to container port 8080. | pass→pass | 4,864 | 7,009 | +44% | 1 | 1 | 0% | 880 | 1,635 | +86% | 0 | 0 | — |
▸case-14 Our pod requires custom ServiceAccount credentials named app-sa to interact with cluster APIs. Someone suggested modifying the default account with cluster-admin rights. Show how to bind the custom service account to the pod spec and disable automatic API token mounting when not needed. | fail→pass | 9,109 | 14,227 | +56% | 1 | 1 | 0% | 1,482 | 2,607 | +76% | 0 | 0 | — |
▸case-15 We want Prometheus scrapers to automatically collect application metrics on port 9090 at path /metrics. Write the pod template metadata annotations required in the deployment YAML to enable automated scraping. | pass→pass | 6,146 | 5,699 | -7% | 1 | 1 | 0% | 963 | 1,390 | +44% | 0 | 0 | — |
▸case-16 We are defining infrastructure using Terraform HCL. An engineer recommended executing local-exec bash commands with kubectl apply inside Terraform. Write a native Terraform resource definition for a deployment named web-app using the official Kubernetes Terraform provider. | pass→pass | 10,954 | 11,553 | +5% | 1 | 1 | 0% | 1,782 | 2,637 | +48% | 0 | 0 | — |
▸case-17 We are packaging our service into a Helm chart. A teammate proposed hardcoding image tags and replica counts directly in templates/deployment.yaml. Provide the Helm deployment template YAML snippet that dynamically evaluates replica count and container image from values file variables. | pass→pass | 8,000 | 8,163 | +2% | 1 | 1 | 0% | 1,259 | 1,929 | +53% | 0 | 0 | — |
▸case-18 We have a base deployment manifest in our repository and need to scale the replica count to 5 specifically for production without duplicating the entire manifest. Write a production kustomization.yaml overlay file that patches the base deployment. | pass→pass | 7,891 | 10,471 | +33% | 1 | 1 | 0% | 1,568 | 1,919 | +22% | 0 | 0 | — |
▸case-19 We need our microservice pods distributed evenly across AWS availability zones in us-east-1. A team member suggested deploying 3 separate deployments with static nodeSelector constraints for each zone. Show how to configure topologySpreadConstraints in a single deployment spec for cross-zone distribution. | pass→pass | 14,393 | 13,304 | -8% | 1 | 1 | 0% | 2,226 | 2,898 | +30% | 0 | 0 | — |
▸case-20 Our cloud load balancer controller requires pods to satisfy an external readiness condition status before receiving live traffic. Show how to configure readinessGates on the deployment pod spec for a custom health condition key. | pass→pass | 9,651 | 16,412 | +70% | 1 | 1 | 0% | 1,609 | 2,798 | +74% | 0 | 0 | — |
▸case-21 A container security audit flag requires stripping unneeded Linux kernel capabilities from our deployed pods. Someone suggested running the pod with privileged: true to resolve permission warnings. Provide the securityContext container block that drops all Linux capabilities and re-adds only NET_BIND_SERVICE. | pass→pass | 6,734 | 5,151 | -24% | 1 | 1 | 0% | 1,271 | 1,252 | -1% | 0 | 0 | — |
▸case-22 In a service mesh enabled namespace, we need to explicitly prevent Istio from injecting sidecar proxies into a high-throughput proxy pod. Provide the pod template metadata annotation required in the deployment spec to opt out of Istio sidecar injection. | pass→pass | 3,555 | 5,763 | +62% | 1 | 1 | 0% | 662 | 1,419 | +114% | 0 | 0 | — |
▸case-23 Our compiled Java application Docker image is over 1 GB because Maven build dependencies are left inside. Write a multi-stage Dockerfile that builds the application using Maven in stage 1 and copies only the JAR artifact into an openjdk runtime stage 2. | pass→pass | 12,241 | 9,493 | -22% | 1 | 1 | 0% | 1,915 | 2,125 | +11% | 0 | 0 | — |
▸case-24 We need to provision an AWS Virtual Private Cloud with 3 public subnets and 3 private subnets across availability zones in us-west-2 using Terraform. Provide the Terraform HCL block using the official terraform-aws-modules/vpc module. | pass→pass | 17,417 | 10,781 | -38% | 1 | 1 | 0% | 1,891 | 2,791 | +48% | 0 | 0 | — |
▸case-25 Create a GitHub Actions workflow YAML file located at .github/workflows/ci.yml that triggers on pull requests to main, checks out the codebase, sets up Node.js 18, and executes npm test. | pass→pass | 4,428 | 4,712 | +6% | 1 | 1 | 0% | 823 | 1,218 | +48% | 0 | 0 | — |