Agentless Monitoring
The agent-based approach to monitoring requires one internal agent per system that is to be monitored. As operating systems and applications have evolved, they have incorporated newer instrumentation mechanisms that allow monitoring of these environments from remote locations - i.e., external to the system or application that is to be monitored. The main advantage of this remote monitoring approach is that it does not require an agent to be installed on every system that is to be monitored - hence, the name “agentless monitoring” for this approach. The table below summarizes the tradeoffs between agentless and agent-based monitoring.
Agentless Monitoring | Agent-based Monitoring |
---|---|
Easier to implement, maintain, and upgrade as there is no need to deploy agents on all the servers |
More difficult to implement as it requires agents on critical servers |
The nature of statistics that can be reported depends on the instrumentation levels and on the access rights provided for monitoring from a remote location |
Typically, used to collect a wealth of availability, performance, usage metrics; can provide tight integration with the operating system and applications monitored |
Mostly used for problem detection and alerting; provides limited information for diagnosis |
Can provide critical information for problem diagnosis and remote control/problem correction |
Significant network overheads; all the monitoring is done over the network |
This configuration optimizes network usage; most of the data collection and analysis is done on the servers and only processed metrics are transmitted; hence, this approach is very efficient in its network usage |
Higher security risk - monitoring from a remote location must be allowed; some ports (e.g., Netbios/SSH) need to be opened for remote access |
No security risk - the agent does not listen on a port, no firewall rules need to be changed, no new ports need to be opened for the monitoring |
Unlike many existing solutions, eG Enterprise does not require that IT administrators choose between one of these contrasting approaches at the time of installing/deploying eG Enterprise. Instead, when adding any new application/system for monitoring, administrators have a choice of whether to use agent-based or agentless monitoring for the application or system under consideration. For instance, an administrator can choose to monitor the most critical servers in an agent-based manner, and to monitor the less critical servers in the staging/development environment in an agentless manner.