▸case-01 Our web service pods in production keep hitting CrashLoopBackOff with OOMKilled status during traffic spikes. Can you give me a structured incident response guide? I need concrete diagnostic commands to inspect cluster resources and container metrics, a step-by-step workflow to isolate memory leaks versus bad limit configurations, immediate mitigation options, and a Prometheus alerting rule snippet to alert us before pods crash. | pass→fail | 27,597 | 25,970 | -6% | 1 | 1 | 0% | 3,839 | 6,291 | +64% | 0 | 0 | — |
▸case-02 We are seeing intermittent 504 Gateway Timeouts from our AWS ALB during peak hours, and distributed tracing via OpenTelemetry shows heavy latency between two internal microservices. Please provide a root-cause investigation plan formatted with actionable triage steps, commands to filter logs and correlate trace IDs across services, emergency stabilization tactics, and a blameless postmortem template. | fail→fail | 40,581 | 35,615 | -12% | 1 | 1 | 0% | 6,648 | 7,519 | +13% | 0 | 0 | — |
▸case-03 Following a recent GitOps deployment via ArgoCD, our application started failing with database connection exhaustion and timeout errors. I need a comprehensive troubleshooting runbook that outlines how to verify the deployment status, commands to inspect connection pool usage and active queries, quick rollback instructions, and preventive monitoring recommendations. | fail→pass | 59,220 | 19,342 | -67% | 1 | 1 | 0% | 2,598 | 5,321 | +105% | 0 | 0 | — |
▸case-04 Applications running in our Kubernetes cluster are failing with intermittent `Name or service not known` errors when attempting to connect to internal services via `.cluster.local` hostnames. We suspect CoreDNS is overloaded or misconfigured. Provide an operational troubleshooting plan with commands to inspect DNS resolution logs, verify CoreDNS deployment endpoints, test DNS queries from inside a debug container, and adjust CoreDNS replica scaling or caching. | pass→pass | 19,199 | 22,445 | +17% | 1 | 1 | 0% | 3,464 | 5,938 | +71% | 0 | 0 | — |
▸case-05 Our Linux production servers are experiencing high network latency and dropped connections under high throughput. System metrics show `netstat -s` reporting increasing TCP SYN backlog drops. We need a system performance debugging guide with `sysctl` kernel parameter tweaks, socket inspection commands, and packet capture instructions. | pass→pass | 23,390 | 24,355 | +4% | 1 | 1 | 0% | 4,178 | 6,360 | +52% | 0 | 0 | — |
▸case-06 A StatefulSet pod in Kubernetes is stuck in `ContainerCreating` with events showing `FailedAttachVolume` and `Multi-Attach error for volume`. The attached cloud persistent disk is stuck attached to a terminated node. Provide a safe troubleshooting guide to identify the lingering volume attachment, detach it safely, and bring the pod back online without risking persistent storage corruption. | fail→fail | 18,235 | 22,604 | +24% | 1 | 1 | 0% | 3,157 | 5,769 | +83% | 0 | 0 | — |
▸case-07 Our primary Redis cache instance is spiking to 100% single-core CPU utilization, leading to application timeout errors. We suspect high-complexity key commands or unbounded key scans are blocking the single-threaded Redis event loop. Provide a troubleshooting workflow including commands to analyze slow queries without blocking production, check memory fragmentation, and mitigate the CPU load. | fail→pass | 19,647 | 18,494 | -6% | 1 | 1 | 0% | 3,301 | 4,942 | +50% | 0 | 0 | — |
▸case-08 Our Apache Kafka event pipeline is accumulating heavy consumer lag on key topics, and the consumer group keeps dropping members in repeated rebalance storms. We need a systematic debugging checklist with Kafka CLI commands to inspect consumer offsets, identify slow record processing, and tune client consumer parameters to prevent frequent rebalances. | pass→pass | 23,747 | 21,795 | -8% | 1 | 1 | 0% | 4,059 | 5,620 | +38% | 0 | 0 | — |
▸case-09 Services inside our Istio service mesh are encountering elevated HTTP 503 Service Unavailable errors and latency overhead. We suspect the Envoy sidecar proxy is failing health checks or failing mTLS handshake negotiations with upstream pods. Provide a step-by-step Envoy sidecar debugging workflow with `istioctl` commands, proxy log level adjustments, and upstream cluster status checks. | pass→pass | 22,767 | 20,614 | -9% | 1 | 1 | 0% | 3,807 | 5,415 | +42% | 0 | 0 | — |
▸case-10 Our PostgreSQL database is reporting repeated `deadlock detected` errors under heavy parallel writes, causing backend REST API requests to fail with HTTP 500 errors. Provide a structured database troubleshooting guide detailing SQL system queries to inspect active lock waits, query cancellation methods, and transaction ordering strategies to prevent deadlocks. | pass→fail | 33,209 | 56,592 | +70% | 1 | 1 | 0% | 6,048 | 3,227 | -47% | 0 | 0 | — |
▸case-11 Our CI/CD pipeline Docker build step is failing intermittently with disk space exhaustion on runner nodes, and successful builds take over 20 minutes due to missing cache hits. Provide a troubleshooting and optimization guide to clean up runner container images safely, restructure Dockerfile instructions for multi-stage caching, and reduce image size. | pass→pass | 33,204 | 20,555 | -38% | 1 | 1 | 0% | 5,817 | 5,261 | -10% | 0 | 0 | — |
▸case-12 A Terraform automated deployment pipeline failed mid-apply, leaving the backend S3/DynamoDB state lock acquired. Subsequent pipeline runs fail with `Error acquiring the state lock`. In addition, manual cloud resource changes have created state drift. Provide an operational playbook for unlocking state safely and reconciling state drift without corrupting backend resources. | fail→fail | 32,153 | 20,606 | -36% | 1 | 1 | 0% | 4,563 | 5,361 | +17% | 0 | 0 | — |
▸case-13 Our RabbitMQ cluster message queue is filling rapidly with unacknowledged messages, and messages are continuously dropping into the Dead Letter Queue (DLQ). We need an operational debugging guide to inspect queue depth, analyze message rejection reasons in the DLQ, and safely replay or re-consume failed messages. | pass→pass | 23,526 | 27,540 | +17% | 1 | 1 | 0% | 4,131 | 6,820 | +65% | 0 | 0 | — |
▸case-14 Serverless backend functions deployed on AWS Lambda are timing out after 15 seconds during API spikes, causing API Gateway 504 responses. Distributed traces reveal that cold starts and VPC ENI attachment delays are contributing significantly to latency. Provide a serverless performance debugging and optimization plan. | pass→pass | 24,989 | 21,010 | -16% | 1 | 1 | 0% | 3,738 | 5,260 | +41% | 0 | 0 | — |
▸case-15 Microservice calls in our cluster are suddenly failing with `curl: (60) SSL certificate problem: certificate has expired` and `x509: certificate signed by unknown authority`. Provide an operational guide to inspect TLS certificate expiration dates, validate intermediate CA bundle chains, and verify endpoint certificates using OpenSSL CLI tools. | pass→pass | 29,321 | 17,674 | -40% | 1 | 1 | 0% | 4,587 | 5,046 | +10% | 0 | 0 | — |
▸case-16 Our Elasticsearch cluster status has flipped from Green to Red due to multiple unassigned primary shards following a node failure. Search queries are failing with 503 errors. Provide a cluster troubleshooting guide using Elasticsearch REST APIs to diagnose why shards are unassigned and safely restore cluster health. | pass→pass | 19,988 | 22,244 | +11% | 1 | 1 | 0% | 3,358 | 5,668 | +69% | 0 | 0 | — |
▸case-17 Kubernetes nodes are failing into `NotReady` status with conditions reporting `KubeletHasDiskPressure`. The root partition on the node host is at 100% usage due to unrotated container log files in `/var/log/pods` and dangling image layers. Provide host-level troubleshooting commands to safely clear log bloat and prune container images without breaking running pods. | fail→fail | 16,129 | 15,252 | -5% | 1 | 1 | 0% | 2,749 | 4,565 | +66% | 0 | 0 | — |
▸case-18 Applications authenticated with HashiCorp Vault using AppRole authentication are experiencing sudden HTTP 403 Forbidden errors when fetching dynamic database credentials. Logs indicate lease expiration and renewal failures. Provide a troubleshooting runbook to inspect token lease status, renew leases, and ensure client token renewal loops function correctly. | pass→fail | 23,688 | 26,622 | +12% | 1 | 1 | 0% | 4,051 | 6,458 | +59% | 0 | 0 | — |
▸case-19 Our Kubernetes ingress layer powered by NGINX Ingress Controller is returning HTTP 502 Bad Gateway responses for requests to an internal API service. Upstream pods appear to be running, but ingress logs report `connect() failed (111: Connection refused) while connecting to upstream`. Provide an ingress troubleshooting checklist to locate the network/service disconnect. | fail→fail | 20,414 | 14,587 | -29% | 1 | 1 | 0% | 3,030 | 4,475 | +48% | 0 | 0 | — |
▸case-20 Fluent Bit log forwarder pods are consuming excessive memory and crashing due to buffer growth while shipping logs to a remote OpenSearch endpoint. The log collector log files show `[warn] [engine] storage buffer full` and HTTP 429 Too Many Requests response errors from the destination. Provide a log pipeline debugging guide to mitigate buffer bloat, handle downstream backpressure, and optimize storage parameters. | pass→fail | 32,473 | 27,060 | -17% | 1 | 1 | 0% | 5,635 | 5,841 | +4% | 0 | 0 | — |
▸case-21 We are building a new e-commerce microservice in Python using FastAPI. Write the full implementation code for the payment processing checkout endpoint, including user credit card tokenization, calling the Stripe API, saving the order state to PostgreSQL, and returning a JSON response. | pass→pass | 21,776 | 22,869 | +5% | 1 | 1 | 0% | 4,556 | 6,625 | +45% | 0 | 0 | — |
▸case-22 We are starting a brand new project and need to write greenfield Terraform HCL modules from scratch to provision a multi-region AWS infrastructure including VPCs, subnets, internet gateways, and route tables for a new client environment. | pass→fail | 22,000 | 3,960 | -82% | 1 | 1 | 0% | 5,162 | 2,445 | -53% | 0 | 0 | — |
▸case-23 Create a standard greenfield GitHub Actions workflow file in YAML that builds a Node.js 20 project, runs `npm test`, and publishes the npm package to GitHub Packages on tagged releases. | pass→pass | 10,728 | 10,924 | +2% | 1 | 1 | 0% | 2,242 | 3,640 | +62% | 0 | 0 | — |