v1.0.0 发布合并回master (#52)

Co-authored-by: sundapeng <sundp@mail.zgclab.edu.cn>
Co-authored-by: xuxt <xuxt@zgclab.edu.cn>
Reviewed-on: #52

src/metric/client-plugins/all-in-one-full/plugins/dcgm-exporter/check_health.sh

@@ -0,0 +1,55 @@
+#!/bin/bash
+
+# DCGM Exporter 健康检查脚本
+# 输出 JSON 格式结果
+
+set -e
+
+# 检查 DCGM Exporter 健康状态
+check_health() {
+    local url="http://localhost:9400"
+    local metrics_url="$url/metrics"
+    local name="dcgm-exporter"
+    local status="unhealth"
+    local reason=""
+    
+    # 检查 curl 是否可用
+    if ! command -v curl &> /dev/null; then
+        reason="curl 命令不可用，无法进行健康检查"
+        echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+        exit 1
+    fi
+    
+    # 测试根路径连接
+    local http_code=$(curl -s -o /dev/null -w "%{http_code}" "$url" 2>/dev/null || echo "000")
+    
+    if [[ "$http_code" == "200" ]]; then
+        # 测试 metrics 端点
+        local metrics_code=$(curl -s -o /dev/null -w "%{http_code}" "$metrics_url" 2>/dev/null || echo "000")
+        
+        if [[ "$metrics_code" == "200" ]]; then
+            status="health"
+            reason="success"
+            echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+            exit 0
+        else
+            reason="Metrics 端点异常 (HTTP $metrics_code)"
+            echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+            exit 1
+        fi
+    else
+        reason="HTTP 服务异常 (HTTP $http_code)，请检查 DCGM Exporter 是否正在运行在端口 9400"
+        echo "{\"name\": \"$name\", \"status\": \"$status\", \"reason\": \"$reason\"}"
+        exit 1
+    fi
+}
+
+# 主函数
+main() {
+    check_health
+}
+
+# 脚本入口
+if [[ "${BASH_SOURCE[0]}" == "${0}" ]]; then
+    main "$@"
+fi