{
  "uid": "qwen-lab-overview",
  "title": "Qwen L20 · 实验总览",
  "schemaVersion": 39,
  "version": 1,
  "tags": [
    "AIBrix",
    "Qwen3.8",
    "L20",
    "light"
  ],
  "style": "light",
  "timezone": "browser",
  "refresh": "10s",
  "time": {
    "from": "now-30m",
    "to": "now"
  },
  "editable": true,
  "panels": [
    {
      "id": 1,
      "type": "text",
      "title": "范围与口径",
      "gridPos": {
        "x": 0,
        "y": 0,
        "w": 24,
        "h": 3
      },
      "options": {
        "mode": "markdown",
        "content": "**Qwen3.8-27B-FP8 · TP1 · L20 · AIBrix v0.7.0。** 独立实验模型，期望范围曾为 1–4，实际未形成 4 个 Ready 副本，节点与其他业务共享。客户端完整流成功才计入 success；Goodput 门槛为 TTFT≤2s、TPOT≤150ms、E2E≤30s。直方图 P95 为估计值，原始请求给出精确分位数。"
      }
    },
    {
      "id": 2,
      "title": "完整流请求率",
      "type": "timeseries",
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "gridPos": {
        "x": 0,
        "y": 3,
        "w": 12,
        "h": 8
      },
      "targets": [
        {
          "refId": "A",
          "expr": "sum by(result)(rate(qwen_lab_requests_total{job=\"qwen-lab-client\",phase=~\"$phase\"}[$__rate_interval]))",
          "legendFormat": "{{result}}",
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          }
        }
      ],
      "fieldConfig": {
        "defaults": {
          "unit": "reqps",
          "custom": {
            "drawStyle": "line",
            "lineWidth": 2,
            "fillOpacity": 6,
            "spanNulls": false
          }
        },
        "overrides": []
      },
      "options": {
        "legend": {
          "displayMode": "table",
          "placement": "bottom",
          "calcs": [
            "lastNotNull",
            "max"
          ]
        },
        "tooltip": {
          "mode": "multi"
        }
      }
    },
    {
      "id": 3,
      "title": "达标请求率 Goodput",
      "type": "timeseries",
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "gridPos": {
        "x": 12,
        "y": 3,
        "w": 12,
        "h": 8
      },
      "targets": [
        {
          "refId": "A",
          "expr": "sum by(tenant)(rate(qwen_lab_good_requests_total{job=\"qwen-lab-client\",phase=~\"$phase\"}[$__rate_interval]))",
          "legendFormat": "租户 {{tenant}}",
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          }
        }
      ],
      "fieldConfig": {
        "defaults": {
          "unit": "reqps",
          "custom": {
            "drawStyle": "line",
            "lineWidth": 2,
            "fillOpacity": 6,
            "spanNulls": false
          }
        },
        "overrides": []
      },
      "options": {
        "legend": {
          "displayMode": "table",
          "placement": "bottom",
          "calcs": [
            "lastNotNull",
            "max"
          ]
        },
        "tooltip": {
          "mode": "multi"
        }
      }
    },
    {
      "id": 4,
      "title": "客户端 TTFT P95",
      "type": "timeseries",
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "gridPos": {
        "x": 0,
        "y": 11,
        "w": 12,
        "h": 8
      },
      "targets": [
        {
          "refId": "A",
          "expr": "histogram_quantile(0.95,sum by(le,tenant)(rate(qwen_lab_ttft_seconds_bucket{job=\"qwen-lab-client\",phase=~\"$phase\"}[$__rate_interval])))",
          "legendFormat": "租户 {{tenant}}",
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          }
        }
      ],
      "fieldConfig": {
        "defaults": {
          "unit": "s",
          "custom": {
            "drawStyle": "line",
            "lineWidth": 2,
            "fillOpacity": 6,
            "spanNulls": false
          }
        },
        "overrides": []
      },
      "options": {
        "legend": {
          "displayMode": "table",
          "placement": "bottom",
          "calcs": [
            "lastNotNull",
            "max"
          ]
        },
        "tooltip": {
          "mode": "multi"
        }
      }
    },
    {
      "id": 5,
      "title": "客户端 TPOT P95",
      "type": "timeseries",
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "gridPos": {
        "x": 12,
        "y": 11,
        "w": 12,
        "h": 8
      },
      "targets": [
        {
          "refId": "A",
          "expr": "histogram_quantile(0.95,sum by(le,tenant)(rate(qwen_lab_tpot_seconds_bucket{job=\"qwen-lab-client\",phase=~\"$phase\"}[$__rate_interval])))",
          "legendFormat": "租户 {{tenant}}",
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          }
        }
      ],
      "fieldConfig": {
        "defaults": {
          "unit": "s",
          "custom": {
            "drawStyle": "line",
            "lineWidth": 2,
            "fillOpacity": 6,
            "spanNulls": false
          }
        },
        "overrides": []
      },
      "options": {
        "legend": {
          "displayMode": "table",
          "placement": "bottom",
          "calcs": [
            "lastNotNull",
            "max"
          ]
        },
        "tooltip": {
          "mode": "multi"
        }
      }
    },
    {
      "id": 6,
      "title": "实际在途请求",
      "type": "timeseries",
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "gridPos": {
        "x": 0,
        "y": 19,
        "w": 12,
        "h": 8
      },
      "targets": [
        {
          "refId": "A",
          "expr": "sum by(tenant)(qwen_lab_inflight{job=\"qwen-lab-client\",phase=~\"$phase\"})",
          "legendFormat": "租户 {{tenant}}",
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          }
        }
      ],
      "fieldConfig": {
        "defaults": {
          "unit": "short",
          "custom": {
            "drawStyle": "line",
            "lineWidth": 2,
            "fillOpacity": 6,
            "spanNulls": false
          }
        },
        "overrides": []
      },
      "options": {
        "legend": {
          "displayMode": "table",
          "placement": "bottom",
          "calcs": [
            "lastNotNull",
            "max"
          ]
        },
        "tooltip": {
          "mode": "multi"
        }
      }
    },
    {
      "id": 7,
      "title": "计划到达与客户端丢弃",
      "type": "timeseries",
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "gridPos": {
        "x": 12,
        "y": 19,
        "w": 12,
        "h": 8
      },
      "targets": [
        {
          "refId": "A",
          "expr": "sum(rate(qwen_lab_scheduled_arrivals_total{job=\"qwen-lab-client\",phase=~\"$phase\"}[$__rate_interval]))",
          "legendFormat": "计划到达",
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          }
        },
        {
          "refId": "B",
          "expr": "sum(rate(qwen_lab_dropped_arrivals_total{job=\"qwen-lab-client\",phase=~\"$phase\"}[$__rate_interval]))",
          "legendFormat": "客户端达到上限",
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          }
        }
      ],
      "fieldConfig": {
        "defaults": {
          "unit": "reqps",
          "custom": {
            "drawStyle": "line",
            "lineWidth": 2,
            "fillOpacity": 6,
            "spanNulls": false
          }
        },
        "overrides": []
      },
      "options": {
        "legend": {
          "displayMode": "table",
          "placement": "bottom",
          "calcs": [
            "lastNotNull",
            "max"
          ]
        },
        "tooltip": {
          "mode": "multi"
        }
      }
    },
    {
      "id": 8,
      "title": "输出 Token 吞吐",
      "type": "timeseries",
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "gridPos": {
        "x": 0,
        "y": 27,
        "w": 12,
        "h": 8
      },
      "targets": [
        {
          "refId": "A",
          "expr": "sum(rate(qwen_lab_output_tokens_total{job=\"qwen-lab-client\",phase=~\"$phase\"}[$__rate_interval]))",
          "legendFormat": "完整响应输出",
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          }
        }
      ],
      "fieldConfig": {
        "defaults": {
          "unit": "suffix:Token/s",
          "custom": {
            "drawStyle": "line",
            "lineWidth": 2,
            "fillOpacity": 6,
            "spanNulls": false
          }
        },
        "overrides": []
      },
      "options": {
        "legend": {
          "displayMode": "table",
          "placement": "bottom",
          "calcs": [
            "lastNotNull",
            "max"
          ]
        },
        "tooltip": {
          "mode": "multi"
        }
      }
    },
    {
      "id": 9,
      "title": "HTTP 与完整流结果",
      "type": "timeseries",
      "datasource": {
        "type": "prometheus",
        "uid": "prometheus"
      },
      "gridPos": {
        "x": 12,
        "y": 27,
        "w": 12,
        "h": 8
      },
      "targets": [
        {
          "refId": "A",
          "expr": "sum by(http_status,result)(qwen_lab_requests_total{job=\"qwen-lab-client\",phase=~\"$phase\"})",
          "legendFormat": "{{http_status}} / {{result}}",
          "datasource": {
            "type": "prometheus",
            "uid": "prometheus"
          }
        }
      ],
      "fieldConfig": {
        "defaults": {
          "unit": "short",
          "custom": {
            "drawStyle": "line",
            "lineWidth": 2,
            "fillOpacity": 6,
            "spanNulls": false
          }
        },
        "overrides": []
      },
      "options": {
        "legend": {
          "displayMode": "table",
          "placement": "bottom",
          "calcs": [
            "lastNotNull",
            "max"
          ]
        },
        "tooltip": {
          "mode": "multi"
        }
      }
    }
  ],
  "links": [
    {
      "title": "实验总览",
      "type": "link",
      "url": "/d/qwen-lab-overview?theme=light",
      "keepTime": true,
      "includeVars": true
    },
    {
      "title": "扩缩容",
      "type": "link",
      "url": "/d/qwen-lab-scaling?theme=light",
      "keepTime": true,
      "includeVars": true
    },
    {
      "title": "推理引擎",
      "type": "link",
      "url": "/d/qwen-lab-engine?theme=light",
      "keepTime": true,
      "includeVars": true
    },
    {
      "title": "路由与入口",
      "type": "link",
      "url": "/d/qwen-lab-routing?theme=light",
      "keepTime": true,
      "includeVars": true
    },
    {
      "title": "GPU 与容量",
      "type": "link",
      "url": "/d/qwen-lab-gpu?theme=light",
      "keepTime": true,
      "includeVars": true
    },
    {
      "title": "控制面",
      "type": "link",
      "url": "/d/qwen-lab-control?theme=light",
      "keepTime": true,
      "includeVars": true
    }
  ],
  "templating": {
    "list": [
      {
        "name": "phase",
        "label": "测试阶段",
        "type": "query",
        "datasource": {
          "type": "prometheus",
          "uid": "prometheus"
        },
        "query": "label_values(qwen_lab_requests_total, phase)",
        "refresh": 1,
        "includeAll": true,
        "allValue": ".*",
        "current": {
          "text": "All",
          "value": "$__all"
        }
      }
    ]
  },
  "annotations": {
    "list": [
      {
        "builtIn": 1,
        "datasource": {
          "type": "grafana",
          "uid": "-- Grafana --"
        },
        "enable": true,
        "hide": true,
        "iconColor": "#2563eb",
        "name": "Annotations & Alerts",
        "type": "dashboard"
      }
    ]
  }
}