{"exercise":"Fetch GET /checks, then compute a cluster health report and POST it to /validate. Group by service, compute healthy_nodes/total_nodes and avg_latency_ms per service, and classify each service: healthy (>75% healthy nodes), degraded (>=50%), unhealthy (<50%). Overall status is the worst service status.","answer_schema":{"overall":"healthy | degraded | unhealthy","services":{"<service_name>":{"status":"healthy | degraded | unhealthy","healthy_nodes":"int","total_nodes":"int","avg_latency_ms":"float (2 decimal places)"}}},"endpoints":{"GET /checks":"raw health check data","POST /validate":"submit your computed answer"}}