The core principles behind agent and agentless are straightforward. The agent approach is the more traditional procedure for data gathering and involves the installation of software (agents) on all computers from which data is required.
Sometimes, the installation process is performed manually for each computer, while other times it is automated via a centralised installation server that pushes software to other computers.
Agentless is actually a bit of a misnomer as all management requires an agent. However, it leverages what’s already available on the server without installing additional software, and its implementation ranges from using built-in SNMP agents to remote shell access.
Those in favor of agentless monitoring often anchor their argument on ease of deployment: you can control everything from one place, there’s no hassle of configuring and updating agents, and without the requirement of a third-party vendor there are fewer headaches associated with licensing and installation. The outcome is a faster time to deployment.
Agentless is also helpful when you need to gather data from various domain managers. Querying these managers provides understanding into workloads, their relationships, and what configuration they’re in, for example. Users are also able to rapidly recognize holes in monitoring coverage.
However, agentless has its drawbacks. While it is feasible to gather logs with an agentless approach, it tends to be pricey and error-prone, and when it comes to collecting performance data, agentless isn’t as bandwidth efficient.
On the other side of the spectrum, we have agent monitoring. The agent approach allows for the collection of more detailed data at higher frequencies with less load on IT infrastructure. It also facilitates the collection of logs generated by applications, often containing important performance metrics and configuration information. And then there’s the issue of security: not only are agents better suited to more secure environments, they also have local rights and can act without external authentication.
However, like agentless, agent monitoring isn’t without its disadvantages. It is necessary to install agents on each server or system in order to begin monitoring which complicates installation and deployment. Some IT teams limit the number of agents they manage because of the potential administrative load. And while agents are great to use in highly dynamic and secure environments, agents lack the visibility to operate outside their own domain.
The complicating factor
The arguments for both approaches has become complicated in recent years. In the old world of owned, static infrastructure, it was relatively easy to discover and monitor infrastructure components. But the ingredients of a modern infrastructure have changed.
Hybrid cloud infrastructure – combining on-premises, private and public clouds – is becoming ever more popular, adopted by businesses looking to leverage the direct benefits of both a private and public cloud environment.
According to research at ScienceLogic, 81 percent of enterprises already have hybrid cloud environments and one third have more than 25 percent of their IT in the cloud. And companies in the cloud are running applications and workloads in more than three clouds, on average,
Hybrid cloud offers the potential to save more on IT costs and to redirect those savings to improved business outcomes, and yet it brings with it some unique challenges. Manual processes have become unmanageable in a hybrid world where cloud instances are spun up and down at will – after all, there are no people in the cloud. The whole agent versus agentless argument suddenly becomes much more complex.
Hybrid cloud has also disrupted configuration management databases (CMDB). These were once an enormously beneficial tool, providing a complete view of all the IT assets in a business. Serving as the single, master list of the IT environment, the CMDB maps out and describes relationships between network infrastructure, systems, servers, applications and, now, virtualised hardware.
But CMDB hasn’t kept up in multi-cloud environments. Some analysts put the percentage of failed and out-of-date CMDB in organizations at 90 percent. It’s no longer the golden source of data it once was.
A new approach
Agents are fantastic for engaging in concentrated, deep, high-fidelity monitoring. They’re also highly applicable to log management, which is an increasingly important facet of modern applications.
Agentless monitoring provides that big picture view of what’s going on in your IT world. It’s the foundation upon which you deploy agents for specific use cases. Agents help you scour more effectively in smaller environments.
When you add the two together, you get a complete picture of what’s happening with your operations, even in a rapidly changing cloud environment – from the application all the way down to the infrastructure. Ensuring an updated CMDB is also vital – it’s important to map out infrastructure. After all, an accurate CMDB is the foundation for coping with changeable cloud environments.
Sourced from Erik Rudin, VP of business development and alliances, ScienceLogic