登录

hertzbeat

apacheApache-2.0活跃维护

开源的实时可观测性系统，通过 Agent 采集指标、日志并接入 LLM 智能分析，替代 Grafana 等传统监控方案做告警通知。

An AI-powered next-generation open source real-time observability system.

v1.9.0发布于 2026-09-13 13:06 UTC7,461star +1 / 7 天1,367fork同步于 2026-11-10 05:32 UTC
可观测性实时监控告警通知指标采集日志监控
对比 GitHub 官网agentaialertingdatabasegrafana

怎么上手

由 README 提炼，执行前请确认
原文 
docker run -d -p 1157:1157 -p 1158:1158 --name hertzbeat apache/hertzbeat
硬件要求按 README 说明

README 未说明显存或 GPU 要求；安装包支持 x86/arm64 CPU 架构，可横向扩展多采集器集群。

有 Docker 镜像
下载：源码 zip·GitHub Releases（安装包）本站不镜像安装包

项目说明

由 README 自动翻译提炼
没人工点评？写下你的经验

Apache HertzBeat 是无需 Agent 的实时可观测性平台，统一采集指标日志并集中告警

开源的实时可观测性系统，通过 Agent 采集指标、日志并接入 LLM 智能分析，替代 Grafana 等传统监控方案做告警通知。

核心能力

  • 集成采集、分析、告警、通知于一体，内置 MCP Server 能力
  • 无 Agent 采集，兼容 Prometheus，覆盖数据库、中间件、云原生等
  • 通过在线配置 YML 模板即可自定义监控类型，如 K8s、Docker
  • 统一日志平台，通过 OTLP 协议接入多来源日志
  • 告警支持分组收敛、静默、抑制，并可分发到邮件、钉钉、Webhook 等渠道

容易踩的坑

源码启动后端需 Java 25 与 maven3+，前端需 Node.js 22 和 pnpm 10，且 IDE 需加 --add-opens 的 VM 参数；默认账号为 admin/hertzbeat，Web 端口 1157、采集器通信端口 1158；完整部署还需另配 PostgreSQL/MySQL 与 VictoriaMetrics/IoTDB/TDengine 等时序库。

内容由模型根据 README 生成，可能不完整或不准确，以官方文档为准。看原文 README

最近发版

5 个版本v1.9.0
全部 release
  1. v1.9.0最新正式版2026-09-13 · 2026-09-13 13:06 UTC

    Note

    This release contains breaking changes 1.9.0 is not a drop-in replacement for the 1.8.x jar or image. The runtime, the configuration file, the authorization rules, the collector credential format, the GreptimeDB table schema and several alerting semantics all changed incompatibly. Read this guide in full before you start: how to update to 1.9.0

    What's Changed

  2. 1.8.0v1.8.02026-01-31 · 2026-01-31 07:12 UTC
    查看更新说明

    Hi guys! Apache HertzBeat 1.8.0 version has published!

    What's Changed

  3. v1.7.32025-08-31 · 2025-08-31 14:27 UTC
    查看更新说明

    Hi guys! Apache HertzBeat 1.7.3 version has published!

    What's Changed

  4. v1.7.22025-06-25 · 2025-06-25 14:17 UTC
    查看更新说明

    v1.7.2

    Hi guys! Apache HertzBeat (incubating) 1.7.2 version has published!

    What's Changed

  5. v1.7.12025-05-29 · 2025-05-29 08:37 UTC
    查看更新说明

    v1.7.1

    Hi guys! Apache HertzBeat (incubating) 1.7.1 version has published!

    What's Changed

用户点评

登录用户的真实使用记录，可按版本和硬件筛选
共 0 条

    还没有人点评。第一个说出真实体验的人最有价值。

    README 原文

    未翻译
    在 GitHub 打开

    Discord Reddit Twitter OpenSSF Best Practices Docker Pulls Artifact Hub Contribute with Gitpod

    Home: hertzbeat.apache.org
    Email: Mail to [email protected] to subscribe mailing lists

    🎡 Introduction

    Apache HertzBeat™ is an AI-powered next-generation open source real-time observability system. Unified metrics and logs collection, centralized alerting distribution, intelligent management and analysis. No Agent required, high performance cluster, provides powerful custom monitoring and status page building capabilities.

    Features

    • Integrates collection + analysis + alerting + notification into one platform, with new AI-powered interactions and features under HertzBeat AI, and built-in MCP Server capabilities.
    • Unified metrics platform, agentless, Prometheus-compatible, supports application services, programs, databases, caches, operating systems, big data, middleware, web servers, cloud-native, networks, custom monitoring and more.
    • Unified logging platform, seamlessly integrates multiple log sources through OTLP protocol for reporting.
    • Unified alerting platform, integrates internal alerts with various external alert sources, unified alert processing and analysis, flexible real-time and periodic threshold rules, grouping convergence, silence, suppression, etc.
    • Unified message distribution, alerts processed by the alerting platform are distributed via Email Discord Slack Telegram DingTalk WeChat FeiShu SMS Webhook ServerChan and other methods.
    • Makes protocols such as Http, Jmx, Ssh, Snmp, Jdbc, Prometheus configurable, allowing you to collect any metrics by simply configuring the template YML file online. Imagine being able to quickly adapt to a new monitoring type like K8s or Docker simply by configuring online with HertzBeat.
    • High performance, supports horizontal expansion of multi-collector clusters, multi-isolated network monitoring and cloud-edge collaboration.
    • Provides powerful status page building capabilities, easily communicate the real-time status of your service to users.

    HertzBeat's unified platform, AI intelligence, powerful customization, multi-type support, high performance, and easy expansion, aims to help users quickly and conveniently achieve observability requirements.

    🥐 Architecture

    hertzBeat

    🐕 Quick Start

    • Community alpha preview users can start from docs/alpha-preview.md for the current Next.js frontend, entity-centered observability, local smoke, and known alpha limitations.

    • If you wish to deploy HertzBeat locally, please refer to the following Deployment Documentation for instructions.

    🍞 Install HertzBeat

    HertzBeat supports installation through source code, docker or package, cpu support x86/arm64.

    1：Install quickly via docker

    1. Just one command to get started

      docker run -d -p 1157:1157 -p 1158:1158 --name hertzbeat apache/hertzbeat

    2. Access http://localhost:1157 to start, default account: admin/hertzbeat

    3. Deploy collector clusters (Optional)

      docker run -d -e IDENTITY=custom-collector-name -e MANAGER_HOST=127.0.0.1 -e MANAGER_PORT=1158 --name hertzbeat-collector apache/hertzbeat-collector
      • -e IDENTITY=custom-collector-name : set the collector unique identity name.
      • -e MODE=public : set the running mode(public or private), public cluster or private cloud-edge.
      • -e MANAGER_HOST=127.0.0.1 : set the main hertzbeat server ip.
      • -e MANAGER_PORT=1158 : set the main hertzbeat server port, default 1158.

    Detailed config refer to Install HertzBeat via Docker

    2：Install via package
    1. Download the release package hertzbeat-xx.tar.gz Download
    2. Configure the HertzBeat configuration yml file hertzbeat/config/application.yml (optional)
    3. Run command $ ./bin/startup.sh or bin/startup.bat
    4. Access http://localhost:1157 to start, default account: admin/hertzbeat
    5. Deploy collector clusters (Optional)
      • Download the release package hertzbeat-collector-xx.tar.gz to new machine Download
      • Configure the collector configuration yml file hertzbeat-collector/config/application.yml: unique identity name, running mode (public or private), hertzbeat manager-host, hertzbeat manager-port 
        collector:
  dispatch:
    entrance:
      netty:
        enabled: true
        identity: ${IDENTITY:}
        mode: ${MODE:public}
        manager-host: ${MANAGER_HOST:127.0.0.1}
        manager-port: ${MANAGER_PORT:1158}
      • Run command $ ./bin/startup.sh or bin/startup.bat
      • Access http://localhost:1157 and you will see the registered new collector in dashboard

    Detailed config refer to Install HertzBeat via Package

    3：Start via source code
    1. Local source code debugging needs to start the back-end project hertzbeat-startup and the front-end project web-app.
    2. Backend: need maven3+, Java 25, and lombok; add VM options in IDE: --add-opens=java.base/java.nio=org.apache.arrow.memory.core,ALL-UNNAMED, then start the hertzbeat-startup service.
    3. Web: need Node.js 22 and pnpm 10, run corepack [email protected] install --frozen-lockfile && corepack [email protected] dev in web-app after backend startup.
    4. Access http://localhost:4200 to start, default account: admin/hertzbeat

    Detailed steps refer to CONTRIBUTING

    4：Install All(hertzbeat+postgresql+tsdb) via Docker-compose

    Install the postgresql/mysql database, victoria-metrics/iotdb/tdengine database and hertzbeat at one time through docker-compose deployment script.

    Detailed steps refer to Install via Docker-Compose

    5: Install All(hertzbeat+collector+postgresql+tsdb) via kubernetes helm charts

    Install HertzBeat cluster in a Kubernetes cluster by Helm chart.

    Detailed steps refer to Artifact Hub

    HAVE FUN

    ✨ Contributors

    Thanks to these wonderful people, welcome to join us: Contributor Guide

    💬 Join discussion

    Channel

    Join the Mailing Lists : Mail to [email protected] to subscribe mailing lists.

    Chat On Discord

    WeChat Group : Add friend ahertzbeat and invite to the group.

    WeChat Public : Search ID usthecom.

    QQ Group : Group num 1035688434

    Github Discussion

    Reddit Community

    Follow Us Twitter

    Subscribe YouTube

    Open-Source Project Build From Open-Source

    HertzBeat is built on so many great open source projects, thanks to them!

    • Java Spring SpringBoot Jpa Maven Assembly Netty Lombok Sureness Protobuf HttpClient Guava SnakeYaml JsonPath ...
    • TypeScript Next.js React Tailwind CSS Node.js Npm ECharts G6 MonacoEditor Docusaurus ...

    Landscape

    🛡️ License

    Apache License, Version 2.0

    相关项目

    同分类里 star 最高的项目
    对比全部 →
    langfuse
    langfuse·TypeScript

    开源 LLM 应用可观测与评估平台，能追踪调用链路、跑评测实验，帮团队定位问题并持续优化提示词和模型效果。

    LLM 可观测+4LLM 可观测
    37224+1004.2kv4.77.02026-11-10 05:09 UTC
    signoz
    SigNoz·TypeScript

    基于 OpenTelemetry 的开源可观测性平台，把日志、指标和链路追踪集中到一处，提供 APM、分布式追踪和基础设施监控，帮团队快速定位线上问题。

    可观测性+4LLM 可观测
    32565+212.6kv0.149.02026-11-10 01:47 UTC
    skywalking
    apache·Java

    分布式系统的应用性能监控平台，支持链路追踪、指标采集和日志分析，帮你在微服务架构下快速定位性能瓶颈。

    应用性能监控+4LLM 可观测
    24996+26.6kv11.1.02026-11-10 04:29 UTC
    jaeger
    jaegertracing·Go

    CNCF 出品的分布式追踪平台，采集服务调用链路并做可视化分析，帮你在微服务和 LLM 应用里定位延迟与故障根因。

    分布式追踪+4LLM 可观测
    23391+53.2kv2.23.02026-11-10 05:16 UTC