EclipseRMM green eclipse symbolEclipseRMM

FIELD NOTES / RMM FUNDAMENTALS

What Is an RMM Agent? How Agent-Based Windows Management Works

What is an RMM agent? It is a small program installed on each computer you manage. It runs in the background, keeps a connection to the remote monitoring and management (RMM) portal, reports that the device is online, and carries out the administrative actions an authorized technician requests. On Windows, the agent usually runs as a Windows service, so it keeps working when nobody is signed in. This guide explains what an RMM agent does, how agent-based management works from install to removal, and what to check before you deploy one.

What is an RMM agent, in plain terms?

Think of the RMM portal as the place where you work and the agent as its representative on each device. You don't open a remote desktop session or sign in to every PC. Instead, you choose a device in the portal and request an action, and the agent on that device performs it and sends the result back.

Three jobs define almost every agent:

Why an RMM agent runs as a Windows service

Microsoft describes a service application as a program that follows the rules of the Service Control Manager (SCM). Microsoft's documentation explains that a service "can be started automatically at system boot" and that services "can execute even when no user is logged on to the system". Those two properties are exactly what a management agent needs: it should be available after a reboot and while the PC sits at the sign-in screen.

The SCM is also what you query when you inspect a service with ordinary Windows tools. You can check any service's state, start type and account yourself:

# Read-only: state and start type of a service
Get-Service -Name Spooler | Format-List Name, DisplayName, Status, StartType

# Read-only: the account the service runs under and its process ID
Get-CimInstance -ClassName Win32_Service -Filter "Name='Spooler'" |
    Format-List Name, State, StartMode, StartName, ProcessId

EclipseRMM's agent follows this model. The EclipseRMM Windows service agent, named EclipseRMMAgent, runs as a Windows service with automatic startup, and normal operation needs neither a persistent desktop window nor a tray application. Administrators can identify and inspect it with ordinary Windows tools, the same way as in the commands above. For a fuller walkthrough of these commands, see how to check service status with PowerShell.

Which account does the agent run as, and why it matters

Every Windows service runs under an account, and that account decides what the service can do. A service can run under LocalSystem. Microsoft's documentation describes LocalSystem as a predefined local account used by the service control manager that "has extensive privileges on the local computer, and acts as the computer on the network". Its token includes the NT AUTHORITY\SYSTEM and BUILTIN\Administrators SIDs.

That matters for how you treat the portal. The EclipseRMM agent service runs as LocalSystem, so authorized management can perform privileged operations on the device. The product pages are direct about the consequence: treat the portal as an administrative interface, use a private login, avoid unreviewed scripts and never operate an endpoint without permission. Whatever agent you evaluate, find out which account it runs as before you install it.

How agent-based management works, step by step

The details vary between products, so here is the lifecycle as it works in EclipseRMM, using only what the product documentation states.

1. Authorize and download

In the portal, you select an organization and explicitly authorize management for devices you own or are permitted to administer. You then download one file: the signed installer ClientUpdater_en_X64.exe for the organization you selected.

2. Install with normal UAC

You run the installer on the authorized PC. Windows User Account Control (UAC) asks for administrative approval. Microsoft explains that with UAC, "each application that requires the administrator access token must prompt the end user for consent". With valid prior management authorization, the EclipseRMM installation then completes silently and starts the service. EclipseRMM does not bypass Defender, SmartScreen or Windows Firewall, and SmartScreen may still prompt. The one-click agent deployment guide walks through the four steps.

3. Enroll and report online

The service enrolls, a permanent device credential is issued during enrollment, and the device appears in the selected organization. From then on, Online status reflects recent authenticated heartbeats.

4. Manage with PowerShell

An authenticated owner with the required permission and an active subscription can request PowerShell operations on an enrolled, authorized device. Execution has time limits, the portal returns output, errors and completion state, and actions are recorded in an audit trail. See PowerShell RMM for how commands and results work.

5. Remove accountably

When the agent should go, the dashboard's confirmed Uninstall action removes the service and revokes device access. Local administrator recovery and removal remain available if dashboard access is unavailable. A later reinstall needs fresh authorized enrollment.

What heartbeats tell you, and what they don't

A heartbeat is a regular check-in from the agent. When heartbeats arrive, the device shows Online; when they stop, the device becomes Offline. In EclipseRMM, an offline device stays in the portal and reconnects with its registered identity when connectivity returns. Offline is not evidence that the agent was uninstalled, and a disconnected device cannot be assumed to have run a request. The guide to RMM heartbeat monitoring and offline status covers how to investigate a device that stopped checking in. If your devices aren't domain-joined, see how a Windows RMM manages workgroup PCs compared with Group Policy and PowerShell remoting.

Is it safe to install an RMM agent?

An RMM agent is privileged software by design, so the honest answer is that its safety depends on where it came from, who can use the portal and how carefully commands are reviewed. A few checks apply to any agent:

A valid signature tells you who published the file. As EclipseRMM's verification page puts it, signing does not eliminate UAC or guarantee SmartScreen reputation, and a hash match checks file identity without replacing signature validation. Treat each check as one piece of evidence, not a guarantee.

RMM agent FAQ

Does an RMM agent need a user to be signed in?

Not when it runs as a Windows service. Microsoft's documentation states that services can run when no user is logged on, and the EclipseRMM agent starts with Windows.

Does the agent show a window or tray icon?

That depends on the product. EclipseRMM's normal operation needs neither a persistent desktop window nor a tray application; the service is visible in ordinary Windows tools such as Get-Service and the Services console.

What happens if the device goes offline?

Heartbeats stop and the portal marks it Offline. In EclipseRMM the device remains in the portal and reconnects when connectivity returns, so you shouldn't reinstall just because a device is temporarily offline.

How many devices can one agent manage?

One. Each managed computer runs its own agent. Pricing models differ between vendors; EclipseRMM Unlimited is $100 USD a month for unlimited authorized Windows devices and multiple organizations.