ELM provides great flexibility and scalability for collecting a variety of health, status and event data from numerous systems. This can be accomplished with different types of agents, however there are certain limitations in the monitoring capabilities of agents you should be aware of when planning your deployment or adding on to an existing installation. In addition to monitoring capabilities, there are other performance factors to consider when either a service agent or virtual agent could be utilized to perform the same monitoring function.
A Service Agent is an executable (TNTAgent.exe) with companion files. The TNTAgent.exe process is installed as the ELM Service Agent service on the monitored computer. It communicates with an ELM Server via TCP sockets and all communications are encrypted. Service Agents can be installed on Windows. All Monitor Items assigned to a Service Agent are executed within the TNTAgent.exe process (except Agent Monitor & Ping Monitor). Only Service Agents can monitor systems in real-time.
- Distributes some of the CPU cycles required to monitor the computer to the computer being monitored. The amount of CPU used by a Service Agent depends on the configured monitor items and their scheduled intervals, and the amount of data to be collected.
- Provides system-level caching. If the ELM Server or network between the ELM Server and Agent computer is unavailable, the Agent process will cache data to the Agent’s local disk drive, then transmit the data when connectivity is restored.
- Provides user defined parameters for setting minimum and maximum cache size and file location.
- Provides single TCP socket port connectivity to the ELM Server, which is necessary if a firewall separates the ELM Server and Agent computer.
- Less network bandwidth is required for communication between the ELM Server and the Service agent on the monitored computer.
- Service Agents must be updated when upgrading ELM.
A Virtual Agent provides remote monitoring. No software is installed on the monitored system. Virtual Agents can be used to monitor Windows computers. All Monitor Items assigned to a Virtual Agent are executed within the ELM Server process running on the ELM Server computer. Many of them use RPC Win32 API calls and therefore require RPC and NetBIOS connectivity between the ELM Server and the Agent.
- Uses less processor CPU on the monitored computer.
- No software is installed on the monitored computer.
- Higher network overhead due to authentication, encryption, and other RPC functions.
- RPC and NetBIOS dependency make them firewall unfriendly
- Because Virtual Agents must run at scheduled intervals, Virtual Agents cannot be monitored in real-time.
- ELM Server service account requires administrative rights on all monitored computers.
IP Virtual Agents
IP Virtual Agents provide remote monitoring of internet services.
- Low cost monitoring of non-windows devices/servers (firewalls, routers, linux) for syslog and snmp.
- Monitoring is limited to certain features
- Typically require longer scheduled intervals
The table below demonstrates the capabilities of each type agent in relation to the monitoring function.
|ELM Monitor Type||Service Agent||Virtual Agent||IP Virtual Agent|
|Event File Collector||Y||Y||–|
|SQL Server Monitor||Y||Y||–|
|Web Page Monitor||Y||–||–|
Note: Certain monitoring functions can also be run at or require running from the ELM Server.
As always, if you have any questions regarding the best agent for the job for monitoring your systems, do not hesitate to contact TNT Software. We’re more than happy to guide you through the process and provide best practice recommendations to aid in your decision making.
We hope that you found this article on Using the Right ELM Agent for Your System Monitoring Needs informative and useful and wish you continued success with ELM.