What problem does it solve? When the Buttercup Cyber Reasoning System misbehaves on Kubernetes, engineers face cascading pod restarts, Redis outages, disk saturation, and stalled task queues with no clear starting point. This Skill provides a structured triage workflow, ready-to-run kubectl and redis-cli commands, and a catalog of known failure patterns to pinpoint root causes quickly. ## Core Features & Use Cases - Guided Triage Workflow: Starts with pod status, events, and warning filters, then narrows down via restart reasons, previous-container logs, and resource inspection. - Redis and Queue Diagnostics: Inspects Redis memory, AOF persistence, client connections, and all 13 Buttercup stream queues plus consumer group lag and task registry state. - Failure Pattern Reference: Documents root causes and fixes for Redis AOF cascades, corpus disk saturation, DinD failures, OOM kills, queue backlogs, health check staleness, and stuck init containers. - Automated Diagnostic Script: Runs a one-shot snapshot collecting pod status, restarts, events, resource usage, Redis health, and queue depths, with an optional full log dump. - Use Case: Multiple pods in the crs namespace suddenly restart together. The Skill guides you to check previous logs across services, identify the shared redis.exceptions.ConnectionError, and debug the Redis AOF fsync stall instead of chasing individual services. ## Quick Start Ask the agent to run the diagnose script and triage why pods in the crs namespace are restarting, then follow the failure pattern reference for the identified symptom.
