Difference Between Agent Based or Agent less

Modern day networks are made up of physical servers, virtual servers, cloud based servers and even legacy servers all running alongside each other.

We will increasingly see the continuous adoption and evolution of Micro services. The network infrastructure today is more complex than ever and monitoring the networks, servers and applications is crucial for various reasons like security, compliance, troubleshooting and more intelligent business decisions.

Before we go ahead with what factors may influence the decision of choosing either of them, we need to know what exactly are agent-based or agentless and what are the differences between them.

Fundamentally you may need both since choosing either one or the other doesn’t fit all. For example Agentless has less overhead but logically it cannot pull data at per second interval from target servers.

In some instance, you would need a one second polling interval e.g. a stock trading application – where each and every spikes of CPU, Memory and Disk I/O counts.

Let’s look at what benefits an agent-based Server monitoring solution has over an agentless with a brief comparison of them.

Ease of Deployment Agent needs to be deployed on every server. Easier to deploy as software installation is required only on the remote data collector.
Data Polling Cycle As low as 1 second
(best suited for I/O sensitive server and application components)
Minimum 60 Seconds on average across the industry
(Greater possibility of losing any critical spike that may occur within 60 second polling cycle window)
Monitoring Technology .Net (PowerShell/WMI)/Python/Java/Go based agents SNMP/PowerShell/WMI/SSH etc.
Security Much more secure than agentless monitoring. The agent to application / OS communications are handled internal to the server. Hence, no additional firewall rules need to be configured. Not as secure because the remote data collector must be allowed to communicate with the target system on different ports. The data collector may also need to be installed with domain administration privileges to be able to access the remote systems.
Network Overheads Very bandwidth efficient since data is collected locally and only the processed final results are transported to the console. So in a way it is one way Push mechanism Introduces additional network traffic as the raw performance data is transported to a remote data collector. It is a two way communication.
Server Overheads Agent gets deployed on target server – while collecting local data, agent will consume CPU cycle as well as low amount of memory. This greatly depends on how low collection frequency is configured. No permanent overhead on the target server. Server CPU cycle is consumed as and when polling request received.
Costs of Implementation Adds up the costs indirectly in terms of more infrastructure that is required to support the agent No additional direct or indirect costs associated
Monitoring Coverage Provides deeper, broader monitoring. Can be limited because not all applications and systems have built-in monitoring capabilities

So it’s a trade off and depends on the requirements and more factors that will help an organisation decide what approach to take.

There are pros and cons of both the approaches and the agent based deployment does not require a substantial investment in resources or costs as such compared to the agent less deployment.

There are hassles in agent based monitoring depending upon how much access is allowed for its installation. In a Government or large scale industry, agent based deployment may be time consuming if there are a lot of approvals and testing required.

Moreover, it might require additional time to maintain agents and lastly the scale of operations might prove as a hassle if the deployment, connection, management and administration is of a huge number of clients and servers.

The final decision on the approach to be adopted should depend on the needs of the client rather than the vendor’s ease of marketing and selling it.

A mix of both agentless and agent-based network monitoring would probably strike the best possible balance of needs, features and costs and could work the best for the company.

As such, there is no best approach between agent-based and agentless network monitoring solutions but if possible an agent based solution should be chosen as it has more data coverage and can help in better network management with scale.

What kind of monitoring tools do you use in your organisation? Share your thoughts on the same with us here.