▸case-06 We are building our deployment pipeline for cloud-native microservices. Our team currently uses Jenkins to run `kubectl apply -f manifests/` directly against production API servers from CI workers upon git merge. Is this pipeline optimal, or should we adopt a continuous deployment mechanism where an agent inside the cluster pulls state from Git? | pass→pass | 17,091 | 20,821 | +22% | 1 | 1 | 0% | 2,780 | 5,407 | +94% | 0 | 0 | — |
▸case-18 Static vulnerability scanning catches known CVEs in container images during build, but fails to detect zero-day attacks, reverse shell executions, or unauthorized file writes inside running containers. How should we monitor runtime container behaviors at the Linux kernel system call level? | pass→pass | 23,301 | 23,843 | +2% | 1 | 1 | 0% | 3,647 | 5,743 | +57% | 0 | 0 | — |
▸case-01 We are planning to move our SaaS backend microservices to AWS EKS across multiple environments. Could you design an enterprise platform architecture and GitOps workflow for us? Please provide your recommendation as a technical proposal structured into four main parts: 1) EKS cluster topology and multi-tenant isolation, 2) GitOps rollout strategy using ArgoCD with canary deployments, 3) Pod security and network policy guardrails, and 4) Observability and alerting stack integration. | fail→fail | 44,608 | 36,171 | -19% | 1 | 1 | 0% | 8,284 | 8,411 | +2% | 0 | 0 | — |
▸case-02 Our GKE production environment's monthly spend has increased significantly due to over-provisioned workloads and fluctuating traffic spikes. Can you formulate a comprehensive K8s FinOps and autoscaling strategy? Please structure the output as an actionable engineering plan that details: workload right-sizing methodology, event-driven and pod/node autoscaler configurations, spot instance strategy with fallback handling, and cost-allocation tooling recommendations. | fail→fail | 40,011 | 38,484 | -4% | 1 | 1 | 0% | 7,218 | 9,011 | +25% | 0 | 0 | — |
▸case-03 I need to set up a local Kubernetes environment on my MacBook for personal testing. I am considering kind or minikube. Can you give me a simple shell script to create a single-node local cluster with kubectl context configured? | pass→pass | 10,807 | 10,883 | +1% | 1 | 1 | 0% | 2,019 | 4,120 | +104% | 0 | 0 | — |
▸case-04 Our Node.js express API service is throwing `TypeError: Cannot read property 'id' of undefined` at user.controller.js line 42 when processing incoming requests in our Kubernetes pod. Here is the controller snippet: `const userId = req.body.user.id;`. How do I fix this JavaScript error? | pass→pass | 9,844 | 6,834 | -31% | 1 | 1 | 0% | 1,802 | 3,199 | +78% | 0 | 0 | — |
▸case-05 We are deploying a small 2-container app (Nginx and PostgreSQL) on a single Ubuntu VM using Docker Compose. We do not use Kubernetes. How should we configure the `docker-compose.yml` file to ensure the database container persists data across restarts? | pass→pass | 12,273 | 11,325 | -8% | 1 | 1 | 0% | 2,326 | 4,038 | +74% | 0 | 0 | — |
▸case-07 We want to implement canary deployments for our production microservices. Currently, we shift traffic by manually updating replica counts between old and new deployments behind a single Service. What modern Kubernetes-native traffic management pattern and progressive delivery approach should we adopt for automated canary rollouts with automated rollback? | pass→pass | 17,911 | 23,602 | +32% | 1 | 1 | 0% | 3,016 | 5,907 | +96% | 0 | 0 | — |
▸case-08 We store application configurations in Git repositories for GitOps deployment. However, our team currently commits base64-encoded Kubernetes Secret manifests directly into the Git repository. What secret management approach should we use to avoid exposing plain secrets while maintaining declarative GitOps synchronization? | pass→pass | 14,961 | 19,077 | +28% | 1 | 1 | 0% | 2,477 | 5,149 | +108% | 0 | 0 | — |
▸case-09 We are upgrading our cluster from legacy PodSecurityPolicies which have been removed. Our developers currently run containerized workloads with root privileges and write access to host root filesystems. What security standard levels should we enforce across namespaces, and how should we configure pod security contexts? | pass→pass | 18,213 | 33,151 | +82% | 1 | 1 | 0% | 3,066 | 4,517 | +47% | 0 | 0 | — |
▸case-10 We have worker pods that process messages from an Apache Kafka topic and AWS SQS queues. Currently, we use the standard Horizontal Pod Autoscaler based on CPU utilization, but during high message bursts, CPU remains low while queue lag spikes, causing processing delays. How should we configure autoscaling for these queue consumers? | fail→pass | 17,335 | 21,339 | +23% | 1 | 1 | 0% | 2,987 | 5,628 | +88% | 0 | 0 | — |
▸case-11 In our shared multi-tenant cluster, development teams currently request `ClusterRoleBinding` grants with `cluster-admin` privileges so their CI pipelines can manage resources across the entire cluster. How should we structure namespace isolation and role-based access controls to enforce proper tenant separation? | pass→pass | 18,369 | 19,609 | +7% | 1 | 1 | 0% | 3,140 | 5,492 | +75% | 0 | 0 | — |
▸case-12 Our Prometheus instance is crashing with OOM errors because we increased local TSDB retention to 12 months on a single volume to satisfy compliance reporting. How should we re-architect our observability stack for multi-cluster, long-term metric storage and query federation? | pass→pass | 22,571 | 26,941 | +19% | 1 | 1 | 0% | 3,672 | 6,395 | +74% | 0 | 0 | — |
▸case-13 Our platform team manages 50+ Kubernetes clusters across multiple clouds using custom Python scripts that invoke cloud vendor CLI commands (`aws eks create-cluster`, `gcloud container clusters create`). How can we standardize cluster provisioning using declarative, Kubernetes-native custom resource definitions and GitOps workflows? | pass→pass | 23,763 | 24,870 | +5% | 1 | 1 | 0% | 3,826 | 6,268 | +64% | 0 | 0 | — |
▸case-14 We are struggling to manage complex traffic routing across multi-tenant teams using legacy Ingress manifests, requiring dozens of vendor-specific NGINX annotations for header matching, traffic splitting, and TLS termination. What official Kubernetes API standard should we adopt to provide clean separation of concerns between infrastructure operators and application developers? | pass→pass | 16,184 | 14,477 | -11% | 1 | 1 | 0% | 2,856 | 4,553 | +59% | 0 | 0 | — |
▸case-15 Our CI/CD pipeline builds container images and pushes them to our registry, but malicious or unvalidated third-party images could be deployed directly into our production cluster. How should we establish supply chain security to ensure only cryptographically signed images with Software Bill of Materials (SBOM) are admitted to the cluster? | pass→pass | 19,107 | 19,072 | -0% | 1 | 1 | 0% | 3,094 | 5,177 | +67% | 0 | 0 | — |
▸case-16 Our Kubernetes cluster utilization is under 15%, yet cluster autoscaler keeps spinning up new nodes because pending pods cannot be scheduled. Pod manifests currently omit CPU and memory requests or set equal request and limit values at huge over-estimates. How should we optimize resource allocation and node bin-packing? | pass→pass | 23,050 | 23,473 | +2% | 1 | 1 | 0% | 3,771 | 5,817 | +54% | 0 | 0 | — |
▸case-17 We need automated mutual TLS (mTLS) encryption and traffic metrics between all microservices. However, our cluster resources are constrained and we cannot afford heavy CPU/memory overhead or complex control planes associated with feature-heavy service meshes. What lightweight service mesh architecture meets these requirements? | pass→pass | 14,876 | 20,684 | +39% | 1 | 1 | 0% | 2,342 | 5,482 | +134% | 0 | 0 | — |
▸case-19 We are organizing our GitOps repository strategy. Should we maintain application source code and deployment manifests in a single unversioned git branch, or separate source code from deployment configuration repos using structured directories or branches for dev, staging, and production? | pass→pass | 17,738 | 17,671 | -0% | 1 | 1 | 0% | 2,991 | 4,688 | +57% | 0 | 0 | — |
▸case-20 We run PostgreSQL and MongoDB databases inside Kubernetes with PersistentVolumeClaims on cloud block storage. In the event of a regional cloud outage or accidental namespace deletion, how should we execute cross-region stateful application backup and disaster recovery? | pass→pass | 20,700 | 27,657 | +34% | 1 | 1 | 0% | 3,484 | 6,493 | +86% | 0 | 0 | — |
▸case-21 We want to reduce cloud compute costs by running stateless microservices on cloud provider Spot instances. However, spot instances can be terminated with a 2-minute notice. How should we configure our node pools, pod disruption budgets, and autoscaler fallback to handle spot terminations without service downtime? | pass→pass | 22,984 | 26,770 | +16% | 1 | 1 | 0% | 3,933 | 6,583 | +67% | 0 | 0 | — |
▸case-22 Our security team wrote 30 new strict admission control rules restricting hostPath mounts, privileged containers, and capabilities. We plan to deploy these directly to production with `enforce` mode to immediately secure the cluster. How should policy-as-code changes be tested and deployed safely? | fail→pass | 15,206 | 18,220 | +20% | 1 | 1 | 0% | 2,350 | 4,861 | +107% | 0 | 0 | — |
▸case-23 Our enterprise operates 100+ Kubernetes clusters across North America, Europe, and Asia. Deploying applications and security policies individually to each cluster API endpoint is causing configuration drift. How should we structure multi-cluster fleet management to ensure consistent policy and app deployment across all clusters? | pass→pass | 21,150 | 23,644 | +12% | 1 | 1 | 0% | 3,528 | 5,955 | +69% | 0 | 0 | — |
▸case-24 We are experiencing high network latency and CPU consumption in iptables routing as our cluster scaled to 5000+ Services. Furthermore, standard IP-based firewall rules cannot provide identity-aware micro-segmentation. What modern CNI architecture leverages kernel eBPF to bypass iptables overhead and enforce layer 7 security policies? | pass→pass | 15,018 | 18,487 | +23% | 1 | 1 | 0% | 2,529 | 5,057 | +100% | 0 | 0 | — |