feat: v0.34 健康探针(Architecture B) + Agent v2.1 + 端口监控 + 告警联系人重构
- v0.31: AlertContact 重构(channel_type下拉+webhook_url替代三分立字段) - v0.32: Agent ProcessTop 增强(cmdline/user/cpu_pct/mem_pct) - v0.33: 端口监控 Tab(监听端口自动发现+down联动告警) - v0.34: 健康探针(Agent本地执行HTTP/TCP/Exec→结果随心跳上报) - Agent新增: ContextSwitch/FD/HealthProbe采集器 - 后端新增: health_probe handler + Agent model扩展 - HostMonitor新增Tab: 健康探针/端口监控
This commit is contained in:
1 parent
e1f75f0bc4
commit
d656520428
25 files changed
+1700
-619
No files matched your search
@@ -0,0 +1,140 @@
|
||||
// Package client 提供 Agent 与 AIOps 后端通信的共享 HTTP 客户端。
|
||||
// Reporter 和 TaskClient 共用此客户端,避免重复实现鉴权、重试、上下文传播。
|
||||
//
|
||||
// 署名 Logan.Li
|
||||
package client
|
||||
|
||||
import (
|
||||
"bytes"
|
||||
"context"
|
||||
"encoding/json"
|
||||
"fmt"
|
||||
"io"
|
||||
"log"
|
||||
"net/http"
|
||||
"strings"
|
||||
"time"
|
||||
)
|
||||
|
||||
const (
|
||||
defaultTimeout = 10 * time.Second
|
||||
maxRetries = 3
|
||||
userAgent = "aiops-agent/2.1"
|
||||
)
|
||||
|
||||
// Client 共享 HTTP 客户端,封装鉴权、路径拼接、重试逻辑。
|
||||
type Client struct {
|
||||
baseURL string
|
||||
token string
|
||||
httpClient *http.Client
|
||||
Logger *log.Logger
|
||||
}
|
||||
|
||||
// New 创建客户端实例。
|
||||
func New(server, token string, logger *log.Logger) *Client {
|
||||
return &Client{
|
||||
baseURL: strings.TrimRight(server, "/"),
|
||||
token: token,
|
||||
httpClient: &http.Client{Timeout: defaultTimeout},
|
||||
Logger: logger,
|
||||
}
|
||||
}
|
||||
|
||||
// Do 执行 HTTP 请求并返回原始响应。调用方负责关闭 resp.Body。
|
||||
func (c *Client) Do(ctx context.Context, method, path string, body []byte) (*http.Response, error) {
|
||||
url := c.baseURL + path
|
||||
|
||||
var lastErr error
|
||||
for attempt := 1; attempt <= maxRetries; attempt++ {
|
||||
var reqBody io.Reader
|
||||
if body != nil {
|
||||
reqBody = bytes.NewReader(body)
|
||||
}
|
||||
|
||||
req, err := http.NewRequestWithContext(ctx, method, url, reqBody)
|
||||
if err != nil {
|
||||
return nil, fmt.Errorf("构造请求失败: %w", err)
|
||||
}
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
req.Header.Set("Accept", "application/json")
|
||||
req.Header.Set("User-Agent", userAgent)
|
||||
if c.token != "" {
|
||||
req.Header.Set("Authorization", "Bearer "+c.token)
|
||||
}
|
||||
|
||||
if err := ctx.Err(); err != nil {
|
||||
return nil, err
|
||||
}
|
||||
|
||||
resp, err := c.httpClient.Do(req)
|
||||
if err != nil {
|
||||
lastErr = err
|
||||
c.Logger.Printf("[http] 请求失败(第%d次) %s %s: %v", attempt, method, path, err)
|
||||
if !sleepWithCtx(ctx, attempt) {
|
||||
return nil, ctx.Err()
|
||||
}
|
||||
continue
|
||||
}
|
||||
|
||||
if resp.StatusCode >= 200 && resp.StatusCode < 300 {
|
||||
return resp, nil
|
||||
}
|
||||
|
||||
// 读取错误响应体
|
||||
errBody, _ := io.ReadAll(io.LimitReader(resp.Body, 1024))
|
||||
resp.Body.Close()
|
||||
|
||||
lastErr = fmt.Errorf("HTTP %d %s %s: %s", resp.StatusCode, method, path, string(errBody))
|
||||
|
||||
// 4xx(除 429)不重试
|
||||
if resp.StatusCode >= 400 && resp.StatusCode < 500 && resp.StatusCode != http.StatusTooManyRequests {
|
||||
return nil, lastErr
|
||||
}
|
||||
c.Logger.Printf("[http] 响应异常(第%d次) %s %s: HTTP %d", attempt, method, path, resp.StatusCode)
|
||||
if !sleepWithCtx(ctx, attempt) {
|
||||
return nil, ctx.Err()
|
||||
}
|
||||
}
|
||||
return nil, lastErr
|
||||
}
|
||||
|
||||
// DoJSON 执行请求并解析 JSON 响应到 result。自动关闭响应体。
|
||||
func (c *Client) DoJSON(ctx context.Context, method, path string, reqBody, result interface{}) error {
|
||||
var bodyBytes []byte
|
||||
if reqBody != nil {
|
||||
var err error
|
||||
bodyBytes, err = json.Marshal(reqBody)
|
||||
if err != nil {
|
||||
return fmt.Errorf("序列化请求体失败: %w", err)
|
||||
}
|
||||
}
|
||||
|
||||
resp, err := c.Do(ctx, method, path, bodyBytes)
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
|
||||
if result != nil {
|
||||
body, err := io.ReadAll(resp.Body)
|
||||
if err != nil {
|
||||
return fmt.Errorf("读取响应体失败: %w", err)
|
||||
}
|
||||
if err := json.Unmarshal(body, result); err != nil {
|
||||
return fmt.Errorf("解析响应失败: %w", err)
|
||||
}
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
// sleepWithCtx 带退避的休眠,可被 context 取消。
|
||||
func sleepWithCtx(ctx context.Context, attempt int) bool {
|
||||
timer := time.NewTimer(time.Duration(attempt) * time.Second)
|
||||
defer timer.Stop()
|
||||
select {
|
||||
case <-timer.C:
|
||||
return true
|
||||
case <-ctx.Done():
|
||||
return false
|
||||
}
|
||||
}
|
||||
Reference in new issue
Block a user