opsx
c983d0e3c1
=== 一、两个线上 bug（172.19.16.8 真机取证）===
1) host 标签是字面量 ${HOSTNAME}
Promtail 不做环境变量展开（除非 -config.expand-env + systemd Environment），
实测 21,825 条日志的 host 全是 "${HOSTNAME}"，无法按主机过滤。
→ 安装时 `hostname -s` 探测真实主机名并渲染进配置，结果存 Collector.ResolvedHostname，
让「预览配置」与线上一致。
2) 系统日志时间戳解析失效
TencentOS 4 的 /var/log/messages 是 ISO8601（2026-07-17T15:38:24.923784+08:00），
旧正则按传统 syslog（Oct 28 18:11:06）写死 → timestamp 阶段匹配不上，用的是采集时间，
回填历史日志时间全错。
→ 安装时探测目标机格式：iso8601 → RFC3339Nano；rfc3164 → Jan _2 15:04:05 + Local；
系统日志路径同时探测（Ubuntu 没有 messages，写死会一条都采不到）。
实测：写一条 10 分钟前的 ISO8601 探针行，Loki 里 loki_ts 与该行自身时间偏差 **0 秒**；
另写一条 2 小时前的被 Loki 以 entry too far behind 拒绝 —— 恰好反证时间戳是按事件时间
解析后推送的（旧实现会用采集时间，反而被接收）。
=== 二、B 档：结构化多 Job 编辑器 ===
后端
- models/scrape_job.go：ScrapeJob + ValidateScrapeJobs
校验 job 名格式、绝对路径、**敏感路径拦截**（/etc/shadow、/etc/sudoers、/root/.ssh、
id_rsa、*.pem…）、正则可编译、job/host/__path__ 保留标签名、数量上限。
- services/promtail_host_config.go：多 job 渲染；多路径展开成多个 static_configs；
pipeline 顺序硬约束 cri → multiline → json → timestamp → drop → limit；
encoding 写在 scrape_config 级（放进 static_configs 里 promtail 直接起不来）。
- services/host_profile.go：探测 distro / init / hostname / 时区 / 架构 / journald
与真实存在的系统日志路径。
- CollectorService：renderConfig 实时渲染、PreviewConfig（只渲染不落库）、
Update 支持 scrapeJobs 并校验、安装结果（hostname/syslog 格式/配置快照）落库。
- 新增 POST /api/v1/collectors/:id/preview → {config, warnings, hostname, syslogFormat}。
- installHost：配置先写 .new + `promtail -check-syntax` 通过才替换，重启后轮询
systemctl is-active，失败把 journalctl 尾部带回前端。
前端
- 采集项结构化编辑器：名称/路径/标签常驻，多行合并 / JSON / 自定义时间戳 / 命中丢弃 /
限流折叠在「高级」里；「预览 YAML」弹窗在保存前看到将要下发的配置。
- 本地即时校验 + 旧数据（只有 logPath）自动迁移成 app + system 两个采集项。
=== 三、验证（host-verify-promtail @172.19.16.8）===
- 预览：host=pre-cex-ops-jenkins（无 ${HOSTNAME}），2 job / 3 条系统日志路径 /
ISO8601 正则 / 多行 + drop + 限流 全部按填写渲染。
- 部署：promtail active、NRestarts=0、-check-syntax 通过，主机侧配置与预览逐行一致。
- Loki：新数据 host 全为 pre-cex-ops-jenkins；job=system 415 条/5min、job=app 40 条/5min。
- 负向 7/7 返回指名到字段的 400：/etc/shadow、/root/.ssh/id_rsa、相对路径、坏正则、
job 名带空格、保留标签名、空路径。
镜像：lghost/lokistack:latest + v0.1 = sha256:26172f51187d
258 lines
10 KiB
Go
258 lines
10 KiB
Go
package routers
|
||
|
||
import (
|
||
"net/http"
|
||
"os"
|
||
"path/filepath"
|
||
"strings"
|
||
"time"
|
||
|
||
"lokistack/controllers"
|
||
"lokistack/middleware"
|
||
|
||
"github.com/gin-contrib/cors"
|
||
"github.com/gin-gonic/gin"
|
||
)
|
||
|
||
func SetupRouter() *gin.Engine {
|
||
// 用 gin.New() 而不是 gin.Default()：Default 自带文本格式的 gin.Logger()，
|
||
// 这里换成结构化 JSON 访问日志（middleware.AccessLog），便于 Promtail→Loki
|
||
// 按字段过滤，同时跳过健康检查/静态资源，避免探针轮询把日志量刷满。
|
||
// Recovery 必须显式加回来（Default 里包含它），否则 panic 会打挂进程。
|
||
r := gin.New()
|
||
r.Use(middleware.AccessLog(), gin.Recovery())
|
||
|
||
// CORS
|
||
cfg := cors.DefaultConfig()
|
||
cfg.AllowAllOrigins = true
|
||
cfg.AllowHeaders = []string{"Origin", "Content-Type", "Accept", "Authorization", "X-Requested-With"}
|
||
cfg.AllowMethods = []string{"GET", "POST", "PUT", "DELETE", "OPTIONS", "PATCH"}
|
||
r.Use(cors.New(cfg))
|
||
|
||
// Controllers
|
||
authCtrl := controllers.NewAuthController()
|
||
clusterCtrl := controllers.NewClusterController()
|
||
hostCtrl := controllers.NewHostController()
|
||
storageCtrl := controllers.NewStorageController()
|
||
deploymentCtrl := controllers.NewDeploymentController()
|
||
collectorCtrl := controllers.NewCollectorController()
|
||
userCtrl := controllers.NewUserController()
|
||
ldapCtrl := controllers.NewLdapController()
|
||
logsCtrl := controllers.NewLogsController()
|
||
alertsCtrl := controllers.NewAlertsController()
|
||
registryCtrl := controllers.NewImageRegistryController()
|
||
|
||
// API v1
|
||
api := r.Group("/api/v1")
|
||
|
||
// 健康检查（公开访问，用于负载均衡和监控）
|
||
api.GET("/health", func(c *gin.Context) {
|
||
c.JSON(http.StatusOK, gin.H{
|
||
"status": "ok",
|
||
"timestamp": time.Now().Unix(),
|
||
"version": "2.0.0",
|
||
})
|
||
})
|
||
|
||
// Prometheus metrics 端点（公开访问，可被Prometheus抓取）
|
||
api.GET("/metrics", func(c *gin.Context) {
|
||
// 基础metrics输出（符合Prometheus文本格式）
|
||
c.Header("Content-Type", "text/plain; version=0.0.4; charset=utf-8")
|
||
c.String(http.StatusOK, "# LokiStack Metrics\n")
|
||
})
|
||
|
||
// Public（登录接口挂 IP 限流防爆破：20 次/分钟）
|
||
api.POST("/auth/login", middleware.LoginRateLimit(20, time.Minute), authCtrl.Login)
|
||
api.POST("/auth/refresh", authCtrl.RefreshToken)
|
||
|
||
// Auth required
|
||
authorized := api.Group("")
|
||
authorized.Use(middleware.JWTAuth())
|
||
|
||
// Auth routes
|
||
authorized.GET("/auth/user", authCtrl.GetUser)
|
||
authorized.POST("/auth/password", authCtrl.ChangePassword)
|
||
|
||
// 日志查询:数据源管理 + Loki 代理查询
|
||
logs := authorized.Group("/logs")
|
||
{
|
||
logs.GET("/sources", logsCtrl.ListSources)
|
||
logs.POST("/sources", middleware.AdminAuth(), logsCtrl.CreateSource)
|
||
logs.PUT("/sources/:id", middleware.AdminAuth(), logsCtrl.UpdateSource)
|
||
logs.DELETE("/sources/:id", middleware.AdminAuth(), logsCtrl.DeleteSource)
|
||
logs.POST("/sources/test", middleware.AdminAuth(), logsCtrl.TestSource)
|
||
logs.GET("/policy-groups", middleware.AdminAuth(), logsCtrl.ListPolicyGroups)
|
||
logs.POST("/policy-groups", middleware.AdminAuth(), logsCtrl.CreatePolicyGroup)
|
||
logs.PUT("/policy-groups/:id", middleware.AdminAuth(), logsCtrl.UpdatePolicyGroup)
|
||
logs.DELETE("/policy-groups/:id", middleware.AdminAuth(), logsCtrl.DeletePolicyGroup)
|
||
logs.POST("/query", logsCtrl.Query)
|
||
logs.POST("/context", logsCtrl.Context)
|
||
logs.GET("/labels", logsCtrl.Labels)
|
||
logs.GET("/label/:name/values", logsCtrl.LabelValues)
|
||
logs.GET("/my-scope", logsCtrl.MyScope)
|
||
logs.POST("/histogram", logsCtrl.Histogram)
|
||
// 总量统计（index/stats）：只读索引，用于算总页数
|
||
logs.POST("/stats", logsCtrl.Stats)
|
||
logs.POST("/metrics", middleware.AdminAuth(), logsCtrl.Metrics)
|
||
logs.GET("/health", logsCtrl.Health)
|
||
}
|
||
|
||
// 镜像源(国内加速/海外/私有仓库)
|
||
authorized.GET("/settings/image-registry", registryCtrl.Get)
|
||
authorized.PUT("/settings/image-registry", middleware.AdminAuth(), registryCtrl.Save)
|
||
|
||
// 告警:飞书通知配置 + 告警规则
|
||
alertNotify := authorized.Group("/alerts")
|
||
{
|
||
alertNotify.GET("/notify-config", alertsCtrl.GetNotifyConfig)
|
||
alertNotify.PUT("/notify-config", middleware.AdminAuth(), alertsCtrl.UpdateNotifyConfig)
|
||
alertNotify.POST("/notify-config/test", middleware.AdminAuth(), alertsCtrl.TestNotifyConfig)
|
||
alertNotify.GET("/rules", alertsCtrl.ListRules)
|
||
alertNotify.POST("/rules", middleware.AdminAuth(), alertsCtrl.CreateRule)
|
||
alertNotify.PUT("/rules/:id", middleware.AdminAuth(), alertsCtrl.UpdateRule)
|
||
alertNotify.DELETE("/rules/:id", middleware.AdminAuth(), alertsCtrl.DeleteRule)
|
||
alertNotify.POST("/check", middleware.AdminAuth(), alertsCtrl.CheckNow)
|
||
alertNotify.GET("/deliveries", alertsCtrl.Deliveries)
|
||
alertNotify.GET("/channels", alertsCtrl.ListChannels)
|
||
alertNotify.POST("/channels", middleware.AdminAuth(), alertsCtrl.SaveChannel)
|
||
alertNotify.PUT("/channels/:id", middleware.AdminAuth(), alertsCtrl.SaveChannel)
|
||
alertNotify.DELETE("/channels/:id", middleware.AdminAuth(), alertsCtrl.DeleteChannel)
|
||
alertNotify.POST("/channels/test", middleware.AdminAuth(), alertsCtrl.TestChannel)
|
||
}
|
||
|
||
// LDAP 配置(admin 读写;查询供用户页判断导入按钮显隐,登录用户即可读)
|
||
authorized.GET("/settings/ldap", middleware.AdminAuth(), ldapCtrl.GetConfig)
|
||
authorized.PUT("/settings/ldap", middleware.AdminAuth(), ldapCtrl.UpdateConfig)
|
||
authorized.POST("/settings/ldap/test", middleware.AdminAuth(), ldapCtrl.TestConnection)
|
||
authorized.POST("/settings/ldap/users", middleware.AdminAuth(), ldapCtrl.ListUsers)
|
||
|
||
// Users（admin 专用）
|
||
users := authorized.Group("/users")
|
||
users.Use(middleware.AdminAuth())
|
||
{
|
||
users.GET("", userCtrl.List)
|
||
users.POST("", userCtrl.Create)
|
||
users.DELETE("/:id", userCtrl.Delete)
|
||
users.POST("/batch-delete", userCtrl.BatchDelete)
|
||
users.POST("/:id/reset-password", userCtrl.ResetPassword)
|
||
users.POST("/batch-import-ldap", userCtrl.BatchImportLDAP)
|
||
}
|
||
// 普通用户仅可访问日志查询及自己的账号接口；管理资源统一在服务端保护。
|
||
authorized.Use(middleware.AdminAuth())
|
||
|
||
// Clusters
|
||
authorized.GET("/clusters", clusterCtrl.List)
|
||
authorized.GET("/clusters/:id", clusterCtrl.Get)
|
||
authorized.POST("/clusters", clusterCtrl.Create)
|
||
authorized.PUT("/clusters/:id", clusterCtrl.Update)
|
||
authorized.DELETE("/clusters/:id", clusterCtrl.Delete)
|
||
authorized.POST("/clusters/:id/validate", clusterCtrl.Validate)
|
||
authorized.GET("/clusters/:id/info", clusterCtrl.GetInfo)
|
||
authorized.GET("/clusters/:id/namespaces", clusterCtrl.Namespaces)
|
||
authorized.GET("/clusters/:id/namespaces/:namespace/deployments", clusterCtrl.Deployments)
|
||
|
||
// Hosts
|
||
authorized.GET("/hosts", hostCtrl.List)
|
||
authorized.GET("/hosts/:id", hostCtrl.Get)
|
||
authorized.POST("/hosts", hostCtrl.Create)
|
||
authorized.PUT("/hosts/:id", hostCtrl.Update)
|
||
authorized.DELETE("/hosts/:id", hostCtrl.Delete)
|
||
authorized.POST("/hosts/:id/validate", hostCtrl.Validate)
|
||
authorized.GET("/hosts/:id/info", hostCtrl.GetInfo)
|
||
|
||
// Storage
|
||
authorized.GET("/storages", storageCtrl.List)
|
||
authorized.GET("/storages/:id", storageCtrl.Get)
|
||
authorized.POST("/storages", storageCtrl.Create)
|
||
authorized.POST("/storages/test", storageCtrl.TestConnection)
|
||
authorized.PUT("/storages/:id", storageCtrl.Update)
|
||
authorized.DELETE("/storages/:id", storageCtrl.Delete)
|
||
authorized.GET("/storages/default", storageCtrl.GetDefault)
|
||
|
||
// Deployments
|
||
authorized.GET("/deployments", deploymentCtrl.List)
|
||
authorized.GET("/deployments/:id", deploymentCtrl.Get)
|
||
authorized.POST("/deployments", deploymentCtrl.Create)
|
||
authorized.PUT("/deployments/:id", deploymentCtrl.Update)
|
||
authorized.POST("/deployments/:id/retention", deploymentCtrl.UpdateRetention)
|
||
authorized.GET("/deployments/:id/config", middleware.AdminAuth(), deploymentCtrl.GetLiveConfig)
|
||
authorized.DELETE("/deployments/:id", deploymentCtrl.Delete)
|
||
authorized.POST("/deployments/:id/deploy", deploymentCtrl.Deploy)
|
||
authorized.POST("/deployments/:id/stop", deploymentCtrl.Stop)
|
||
authorized.POST("/deployments/:id/restart", deploymentCtrl.Restart)
|
||
authorized.GET("/deployments/:id/logs", deploymentCtrl.GetLogs)
|
||
// 资源用量（CPU/内存/限流），页面用来判断是否要扩容
|
||
authorized.GET("/deployments/:id/usage", deploymentCtrl.GetUsage)
|
||
authorized.GET("/deployments/stats", deploymentCtrl.GetStats)
|
||
authorized.GET("/deployments/preflight", deploymentCtrl.Preflight)
|
||
authorized.GET("/deployments/:id/ready", deploymentCtrl.Ready)
|
||
authorized.GET("/deployments/versions", deploymentCtrl.GetVersions)
|
||
|
||
// Collectors
|
||
authorized.GET("/collectors", collectorCtrl.List)
|
||
authorized.GET("/collectors/:id", collectorCtrl.Get)
|
||
authorized.POST("/collectors", collectorCtrl.Create)
|
||
authorized.PUT("/collectors/:id", collectorCtrl.Update)
|
||
authorized.DELETE("/collectors/:id", collectorCtrl.Delete)
|
||
authorized.POST("/collectors/:id/install", collectorCtrl.Install)
|
||
authorized.POST("/collectors/:id/uninstall", collectorCtrl.Uninstall)
|
||
authorized.POST("/collectors/:id/preview", collectorCtrl.Preview)
|
||
|
||
// Static files - use executable directory as base
|
||
distPath := findDistPath()
|
||
|
||
if distPath != "" {
|
||
r.Static("/assets", filepath.Join(distPath, "assets"))
|
||
|
||
// favicon：NoRoute 会把所有非 /api/ 请求回退到 index.html，
|
||
// 故此处显式注册 /favicon.svg 与 /favicon.ico，否则浏览器请求图标只会拿到 HTML。
|
||
faviconSVG := filepath.Join(distPath, "favicon.svg")
|
||
serveFavicon := func(c *gin.Context) {
|
||
if _, err := os.Stat(faviconSVG); err != nil {
|
||
c.Status(404)
|
||
return
|
||
}
|
||
c.Header("Content-Type", "image/svg+xml")
|
||
c.File(faviconSVG)
|
||
}
|
||
r.GET("/favicon.svg", serveFavicon)
|
||
r.GET("/favicon.ico", serveFavicon)
|
||
|
||
r.NoRoute(func(c *gin.Context) {
|
||
if strings.HasPrefix(c.Request.URL.Path, "/api/") {
|
||
c.JSON(404, gin.H{"code": 404, "message": "API endpoint not found"})
|
||
return
|
||
}
|
||
c.File(filepath.Join(distPath, "index.html"))
|
||
})
|
||
}
|
||
|
||
return r
|
||
}
|
||
|
||
// findDistPath finds the frontend dist directory
|
||
func findDistPath() string {
|
||
// Try relative to executable
|
||
if ex, err := os.Executable(); err == nil {
|
||
execDir := filepath.Dir(ex)
|
||
// If running via "go run", executable is in temp dir - fall through
|
||
if !strings.Contains(execDir, "/tmp/go-build") {
|
||
for _, rel := range []string{"dist", "../dist", "../../dist"} {
|
||
p := filepath.Join(execDir, rel)
|
||
if _, err := os.Stat(filepath.Join(p, "index.html")); err == nil {
|
||
return p
|
||
}
|
||
}
|
||
}
|
||
}
|
||
// Try working directory relative paths
|
||
for _, p := range []string{"./dist", "../dist", "/app/dist"} {
|
||
if abs, err := filepath.Abs(p); err == nil {
|
||
if _, err := os.Stat(filepath.Join(abs, "index.html")); err == nil {
|
||
return abs
|
||
}
|
||
}
|
||
}
|
||
return ""
|
||
}
|