Agents
Overview
An agent is a lightweight executable deployed on your infrastructure that executes monitoring checks on behalf of the Onagre platform. Agents connect to the Hub — Onagre's central orchestration service — via a permanent, encrypted gRPC connection.
Agents are intentionally simple: they do not make scheduling decisions, store configuration, or act autonomously. The Hub tells each agent what to check, when to check it, and the agent returns the result. This design keeps agents easy to deploy, update, and replace.
Agent types
Onagre offers three categories of agents, depending on where they run and what they monitor.
Public agents
Public agents are managed and hosted by Onagre in cloud infrastructure. They are available to all users by default and can monitor any resource accessible from the public internet — websites, APIs, DNS records, certificates, and more.
- No installation required.
- Multiple agents provide redundancy and geographic distribution.
- Ideal for monitoring external-facing services.
Private agents — Core
Core agents are deployed on your own infrastructure to monitor the host they run on. They automatically collect system-level metrics — CPU, memory, disk, uptime, and load average — without any sensor configuration.
On first connection, two sensors are created automatically:
- Heartbeat — confirms the agent is alive and connected.
- Server — collects and reports system metrics.
Core agents are available starting from the Pro plan. Learn more →
Private agents — Advanced
Advanced agents include everything in Core and unlock all 14 sensor types — HTTP, databases, certificates, network checks, file integrity, and more. They are designed for organizations needing deep, full-stack observability of internal resources.
Advanced agents also support:
- Network Discovery — scan the local network to find hosts and open ports, then create sensors from the results. Learn more →
- Agent load monitoring — track how busy each agent is to plan capacity.
Advanced agents are available exclusively with the Enterprise plan. Learn more →
Comparison
| Public | Core | Advanced | |
|---|---|---|---|
| Hosted by | Onagre | You | You |
| Plan | All plans | Pro+ | Enterprise |
| System metrics | ✗ | ✓ | ✓ |
| Auto-created sensors | ✗ | ✓ | ✓ |
| All 14 sensor types | ✓ (public only) | ✗ | ✓ |
| Network Discovery | ✗ | ✗ | ✓ |
| Agent load monitoring | ✗ | ✗ | ✓ |
| Best for | External resources | Host health | Full-stack monitoring |
Security
- TLS encryption — All communications between agents and the platform are encrypted end-to-end.
- No local secrets — Credentials are managed centrally in the platform through Environments and delivered to agents at execution time. Nothing is stored on disk.
- Minimal footprint — Agents transmit only monitoring results. They never expose internal network details beyond what your sensors explicitly check.
- Authentication — Each agent authenticates with a unique code and API key pair.
Deployment
Private agents (Core and Advanced) are available for Linux and Windows, with support for Docker and Ansible automation.
Summary
| Agent | Where it runs | What it monitors | Plan |
|---|---|---|---|
| Public | Onagre cloud | Public resources (websites, APIs, DNS, certificates) | All |
| Core | Your infrastructure | Host metrics (CPU, memory, disk, uptime) | Pro+ |
| Advanced | Your infrastructure | Everything — full-stack with all 14 sensor types | Enterprise |