Hertzbeat
An AI-powered next-generation open source real-time observability system.
Last verified:
What is Hertzbeat?
Apache HertzBeat is an AI-powered open source real-time observability system that provides unified metrics and logs collection, centralized alerting distribution, and intelligent management and analysis. It is an agentless monitoring system that supports monitoring web services, databases, caching systems, operating systems, middleware, big data platforms, cloud-native environments, and networks without requiring any agent installation on monitored hosts.
Key features include an all-in-one platform integrating collection, analysis, alerting, and notification with AI-powered interactions through HertzBeat AI and built-in MCP Server capabilities. The system supports configurable protocols including HTTP, JMX, SSH, SNMP, JDBC, and Prometheus, allowing users to define custom metrics with simple online YAML configuration. It offers high-performance architecture with horizontal scaling for collector clusters, multi-network isolation, and cloud-edge collaboration capabilities.
HertzBeat includes powerful status page building capabilities to communicate real-time service status to users, similar to GitHub's status page. The system provides flexible alerting rules with timely notifications via Discord, Slack, Telegram, Email, DingTalk, WeChat, FeiShu, SMS, Webhook, and ServerChan. It features alarm grouping, convergence, silence, and suppression to reduce alarm storms.
The tool is designed for DevOps teams, system administrators, SREs, and developers who need comprehensive monitoring without agent deployment overhead. It is suitable for organizations monitoring thousands of servers worldwide, particularly those working with cloud-native environments, microservices, databases, and middleware. The Apache 2.0 open source license makes it accessible for both personal and enterprise use without licensing restrictions.
Hertzbeat pricing
Pricing model: Freemium
Apache HertzBeat is completely free and open source under the Apache 2.0 license. There are no subscription fees, licensing costs, or paid tiers. The software has no monitoring number limits, no monitoring type restrictions, and no pseudo-open source limitations. Users can deploy it on thousands of servers worldwide without additional costs. The only costs involved are infrastructure costs for hosting the system yourself (server hardware, cloud VMs, or Docker containers). Docker deployment: docker run -d -p 1157:1157 apache/hertzbeat with default credentials admin/hertzbeat.
Hertzbeat pros
- Agentless architecture - no agent installation required on monitored hosts
- AI-powered monitoring with HertzBeat AI and built-in MCP Server capabilities
- Complete all-in-one platform: collection + analysis + alerting + notification
- Supports 50+ built-in monitoring types including databases, OS, middleware, cloud-native
- Configurable protocols: HTTP, JMX, SSH, SNMP, JDBC, Prometheus
- Custom monitoring with simple online YAML template configuration
- High-performance horizontal scaling with collector cluster support
- Cloud-edge collaboration for multi-isolated network monitoring
- Powerful status page building capabilities for public service status
- Flexible alerting with grouping, convergence, silence, and suppression
- 10+ notification channels: Discord, Slack, Telegram, Email, Webhook, SMS, and more
- Prometheus-compatible for unified metrics platform
- Unified logging platform with OTLP protocol support
- Full web-based operations with point-and-click UI interface
- Enterprise-grade security with end-to-end encryption for sensitive data
- Apache 2.0 open source with no monitoring number or License restrictions
- Easy Docker deployment with single command: docker run -d -p 1157:1157 apache/hertzbeat
- Alarm integration with third-party systems: Prometheus, Skywalking, AlertManager
- Built-in MCP Server capabilities for AI interactions
- Supports monitoring Kubernetes, Docker, and cloud-native environments
Hertzbeat cons
- Monitoring history charts require external time-series database configuration
- Package deployment requires root permission for ping connectivity monitoring (ICMP)
- Single collector has task limit based on CPU cores (max threads = CPU cores × 16)
- Docker bridge mode cannot access localhost for dependent databases
- TDengine configuration may need adjustment for SQL statement length limits
- Token authentication with Bearer tokens may have URL escaping issues
- No built-in multi-tenancy support yet (requested feature)
- SNMP trap support not yet implemented (requested feature)
- WebUI requires external database (MySQL or TDengine) for full functionality
- Some bugs reported with SSL certificate monitoring data accuracy
Frequently asked questions about Hertzbeat
What is Apache HertzBeat?
Apache HertzBeat is an AI-powered next-generation open source real-time observability system. It is an agentless monitoring system that provides unified metrics and logs collection, centralized alerting distribution, and intelligent management and analysis. It supports monitoring web services, databases, operating systems, middleware, big data, cloud-native environments, and networks without requiring agent installation.
Does HertzBeat require installing agents on monitored hosts?
No, HertzBeat is completely agentless. It uses different protocols (HTTP, JMX, SSH, SNMP, JDBC, Prometheus) to connect directly to end systems and pulls data in PULL form. Users only need to fill in IP, port, account, password, and other parameters in the WEB interface. Sensitive information like passwords is encrypted on all links.
What monitoring types does HertzBeat support?
HertzBeat supports 50+ built-in monitoring types including: Website, HTTP API, Ping, Port Telnet, SSL Certificate, DNS, Nginx, JVM, SpringBoot2/3, MySQL, PostgreSQL, MariaDB, Redis, ElasticSearch, SQL Server, Oracle, MongoDB, Linux, Windows, CentOS, Ubuntu, Tomcat, Kafka, RabbitMQ, Zookeeper, Kubernetes, Docker, and network switches from Cisco, Huawei, H3C, HP, and TP-Link.
How do I deploy HertzBeat?
The easiest way is Docker: docker run -d -p 1157:1157 --name hertzbeat apache/hertzbeat. Then access http://localhost:1157 with default credentials admin/hertzbeat. For full functionality, you can also deploy with external MySQL and TDengine databases. Package deployment is also available but requires root permission for ICMP ping monitoring.
What notification channels does HertzBeat support?
HertzBeat supports Email, Discord, Slack, Telegram, DingTalk, WeChat, FeiShu (Lark), SMS, Webhook, and ServerChan. Notification policies support tag matching and alert level matching to assign different alerts to different receivers. Custom notification templates are also supported for personalized content formatting.
How does custom monitoring work in HertzBeat?
Custom monitoring uses YAML monitoring templates. Users can define monitoring types, metrics, protocols, and parsing methods by configuring a YML file in the browser. All built-in monitoring types are also mapped to YML templates. You can instantly adapt to new monitoring types like K8s or Docker just by configuring online YAML without code changes.
What is the collector cluster feature?
Collector clusters enable horizontal scaling for large-scale monitoring. Multiple collector nodes support automatic load balancing and automatic failover. Monitoring tasks self-schedule in the cluster - if one collector hangs, tasks automatically migrate to other nodes. New collector nodes automatically share collection pressure. Switching between standalone and cluster mode requires no additional component deployment.
How does alerting work in HertzBeat?
HertzBeat supports real-time thresholds (trigger immediately when data is collected) and scheduled thresholds (PromQL expressions calculated within time periods). Alerting features include grouping convergence (deduplication), alarm inhibition (high-level suppresses low-level), and alarm silence (blocks alerts during maintenance). Threshold rules support visual page configuration or flexible expression rules.
What is the status page feature?
Status page capability lets you quickly build an external status page for your product to communicate real-time service status to users, similar to githubstatus.com. Status page component status syncs with monitoring status, and includes fault event maintenance management. This improves transparency, professionalism, user trust, and reduces communication costs.
Is there a limit to how many monitors a single collector can handle?
Yes, the limit depends on CPU cores. Maximum thread count = CPU cores × 16. For example, an 8-core CPU can handle 128 simultaneous collection tasks. Exceeding this shows 'worker pool full' error. Solution: add a new collector in public mode and HertzBeat will automatically distribute tasks to other collectors.