{
  "context": "Tests that the agent sets a numeric Retry_Limit (not False), adds storage.total_limit_size to cap disk buffer growth, enables the HTTP health server, and explains the reliability decisions in accompanying notes.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Retry_Limit numeric",
      "description": "The Kafka OUTPUT includes a Retry_Limit that is a number (e.g., 3, 5) — NOT Retry_Limit False",
      "max_score": 20
    },
    {
      "name": "storage.total_limit_size set",
      "description": "The OUTPUT or SERVICE section includes storage.total_limit_size to cap disk buffer size",
      "max_score": 20
    },
    {
      "name": "HTTP_Server On",
      "description": "The SERVICE section includes HTTP_Server On",
      "max_score": 15
    },
    {
      "name": "HTTP health endpoint documented",
      "description": "reliability-notes.md (or equivalent) references the /api/v1/health endpoint or port 2020 for Kubernetes liveness probe configuration",
      "max_score": 10
    },
    {
      "name": "Reliability measures explained",
      "description": "reliability-notes.md explains how Retry_Limit bounds retries and how storage.total_limit_size prevents disk exhaustion",
      "max_score": 10
    },
    {
      "name": "Mem_Buf_Limit on input",
      "description": "The tail INPUT plugin includes Mem_Buf_Limit",
      "max_score": 10
    },
    {
      "name": "Exclude_Path on input",
      "description": "The tail INPUT plugin excludes fluent-bit's own container logs via Exclude_Path",
      "max_score": 5
    },
    {
      "name": "DB path on input",
      "description": "The tail INPUT plugin includes a DB setting for position tracking",
      "max_score": 5
    },
    {
      "name": "storage.metrics on",
      "description": "The SERVICE section includes storage.metrics on to enable buffer monitoring",
      "max_score": 5
    }
  ]
}