HertzBeat — Agentless Real-Time Observability System

Introduction

HertzBeat is an Apache-incubating open-source real-time monitoring and observability platform. Its defining feature is agentless monitoring: it connects directly to targets via protocols like HTTP, SNMP, JDBC, SSH, and JMX without requiring any agent installation on monitored hosts. It covers infrastructure, databases, middleware, and application monitoring in a single deployment.

What HertzBeat Does

• Monitors servers, databases, APIs, websites, and middleware without installing agents
• Supports 200+ monitoring types out of the box (MySQL, Redis, Nginx, Docker, K8s, etc.)
• Provides configurable alerting with threshold rules and notification channels
• Collects metrics, logs, and traces into a unified dashboard
• Offers a YAML-based monitoring template system for adding custom monitoring types

Architecture Overview

HertzBeat runs as a single Java application with an embedded collector engine. The collector dispatches protocol-specific tasks (HTTP, SNMP, SSH, JDBC, JMX) to gather metrics from targets at configurable intervals. Collected data is stored in a time-series database (IoTDB, TDengine, GreptimeDB, or built-in H2). A Spring Boot web layer serves the management UI and REST API.

Self-Hosting & Configuration

• Deploy via Docker, Docker Compose, or Kubernetes Helm chart
• No external dependencies required for basic use (embedded H2 database and file storage)
• For production, configure an external TSDB (IoTDB, TDengine, or GreptimeDB) for metric persistence
• Add monitoring targets through the web UI or via REST API batch import
• Create custom monitoring types by writing YAML template definitions

Key Features

• True agentless design: monitors via standard protocols with no software on targets
• YAML-driven extensibility for defining new monitoring types without code changes
• Built-in alert center with support for email, webhook, Slack, Telegram, and DingTalk
• Cluster mode for distributed collection across large environments
• Lightweight single-binary deployment with minimal resource requirements

Comparison with Similar Tools

• Zabbix — powerful but requires agents for deep monitoring; HertzBeat is fully agentless
• Prometheus + Grafana — requires exporters on each target; HertzBeat collects directly
• Datadog — commercial SaaS with agents; HertzBeat is self-hosted and free
• Uptime Kuma — focuses on uptime checks; HertzBeat provides full metrics collection across protocols

FAQ

Q: How does agentless monitoring work? A: HertzBeat connects to targets using standard protocols (HTTP, SSH, SNMP, JDBC, JMX) from a central collector. No software needs to be installed on the monitored host.

Q: Can I monitor custom applications? A: Yes. You can define custom monitoring templates in YAML that specify which protocol to use, what metrics to collect, and how to parse the response.

Q: What time-series databases are supported? A: HertzBeat supports IoTDB, TDengine, GreptimeDB, and an embedded H2 database for small deployments.

Q: Is it production-ready? A: HertzBeat is an Apache Incubating project with active development and community support. Cluster mode is available for scaling in production environments.

Sources

• https://github.com/apache/hertzbeat
• https://hertzbeat.apache.org