Understand the key differences between agent-based and agentless monitoring in Cloudmon, so you can choose the right method for each device in your environment.
Cloudmon supports two fundamentally different approaches to monitoring: agent-based and agentless. Both methods collect performance, availability, and health data from monitored devices, but they differ in how they connect, how frequently they collect data, and the depth of visibility they provide. Choosing the right method for each device type is one of the most important decisions when deploying Cloudmon.
In most environments, both methods are used together. Agent-based monitoring is deployed on critical servers, desktops, laptops, and virtual machines where deep visibility and fast detection are required. Agentless monitoring using SNMP or WMI is used for network devices, less critical servers, and any device where installing an agent is impractical or unnecessary.
| Capability | Agent-Based | Agentless (SNMP / WMI) |
| Downtime detection speed | Within 25 seconds via liveness monitoring | Detected at the next polling interval (typically minutes) |
| Minimum polling interval | 1 second (configurable per device) | Longer intervals, dependent on SNMP or WMI polling configuration |
| Software installation required | Yes, Cloudmon agent installed on the device | No, it uses SNMP or WMI protocols via a probe |
| Process and service monitoring | Full process and service monitoring with per-process CPU, memory, PID, and installation path | Limited. WMI supports some process visibility on Windows; SNMP does not |
| Installed software inventory | Full software inventory visible in the Installed Software tab | Not available |
| Automated remediation | Supported. Runbooks can be executed automatically on the agent when an alarm fires | Runbooks can be run on the probe or controller, not on the monitored device directly |
| Digital Experience Monitoring (DEM) | Supported. DEM requires the agent to measure application experience from the user's device | Not available |
| Hyper-V monitoring | Supported. Hyper-V can only be monitored via the Cloudmon agent | Not available |
| Cloud server support | Supported. The agent can be installed on cloud instances (AWS EC2, Azure VM, etc.) | Supported for SNMP-enabled cloud appliances; WMI requires probe connectivity |
| Scalability | Requires individual installation on each device; most suitable for critical devices | Scales easily to large device counts without installation; ideal for broad coverage |
| Resource overhead on monitored device | Minimal. The agent uses a small amount of CPU and memory on the host | None, as no software runs on the monitored device |
| Supported platforms | Windows, Linux (CentOS, Ubuntu, Debian, Fedora, Red Hat, SUSE, Amazon Linux), macOS, Android | Any SNMP-enabled device; WMI for Windows servers only |
Agent-based monitoring is the right choice when you need the highest level of visibility, the fastest possible detection, or features that are only available with an agent installed on the device.
Critical servers and infrastructure: For servers where downtime has significant business consequences, such as database servers, application servers, and domain controllers, agent-based monitoring detects outages and reboots within 25 seconds via liveness monitoring. This is far faster than agentless polling, which detects issues only at the next polling interval. Financial institutions, healthcare providers, and e-commerce platforms running time-sensitive workloads benefit most from this speed.
High-frequency monitoring requirements: The Cloudmon agent supports polling intervals as short as 1 second, providing real-time visibility into CPU, memory, disk, and network metrics. This granularity is not achievable with SNMP or WMI polling. IT admins can enable 1-second polling for critical servers and use longer intervals of up to 1 minute for less critical devices, balancing detail with resource efficiency.
Automated remediation: When a threshold is breached on an agent-monitored device, Cloudmon can automatically execute a remediation script on that device without manual intervention. For example, a high CPU alarm can trigger a script to kill the offending process or restart a service. This automated remediation capability is only available on agent-monitored devices, as agentless monitoring can only run scripts on the probe or controller.
Process, service, and software visibility: The agent provides full visibility into all running processes with per-process CPU and memory consumption, all installed software, and all active Windows services. This level of detail is essential for compliance auditing, capacity planning, and diagnosing performance issues at the application layer. None of this is available through SNMP or WMI at the same depth.
Digital Experience Monitoring: DEM requires the Cloudmon agent to be installed on end-user devices such as laptops and desktops. The agent measures application experience from the user's perspective across device, LAN, WAN, ISP, and application layers. There is no agentless equivalent for DEM.
Hyper-V environments: Hyper-V servers can only be monitored using the Cloudmon agent. If you run Hyper-V virtualization in your environment, agent-based monitoring is mandatory for those hosts.
Agentless monitoring is the right choice when you need broad coverage across many devices, when installing software on target devices is impractical, or when the monitored devices do not support agent installation.
Network devices: Routers, switches, firewalls, and other network infrastructure devices do not support agent installation. SNMP is the standard protocol for monitoring these devices, giving visibility into availability, interface traffic, CPU, memory, and device-specific metrics through over 4,500 built-in templates in Cloudmon.
Large-scale deployments: In environments with hundreds or thousands of devices, agentless monitoring is more practical and cost-effective than deploying agents individually on every device. SNMP credentials are added once and applied across many devices via templates, making it straightforward to scale coverage quickly.
Less critical desktops and laptops: For devices where the depth of visibility provided by an agent is not required, SNMP monitoring provides the essential availability and performance metrics without the overhead of agent installation and maintenance.
Windows servers via WMI: For Windows servers where an agent cannot be installed, WMI provides a rich set of performance metrics including CPU, memory, disk, and network, plus support for 12 Windows application templates covering Active Directory, DHCP, IIS, SQL Server, Exchange, and more. WMI monitoring requires a Windows probe and a domain administrator credential.