Introduction
An agent is a low-footprint, secure application that runs on managed servers. The agent runs as two OS processes: Agent and Agent Shield.
- On Windows, the Agent and Agent Shield run as Windows services.
- On Linux, the Agent and Agent Shield run as OS processes.
Connectivity methods
Agents can be configured to connect to OpsRamp in one of the following ways:
- Connect directly
- Connect using proxy
Connecting directly
Each Agent connects to OpsRamp directly over the internet. The direct connection configuration is suitable for small environments that do not have a local gateway deployed.
Connecting using proxy
Each Agent connects to OpsRamp using a proxy. The gateway includes an embedded proxy and is the recommended proxy for this configuration. The proxy connection configuration is recommended for large environments. When you connect the Agent using a proxy, you avoid opening outbound access for each agent.
Outbound connectivity requirements
| Requirement | Purpose |
|---|---|
| TCP port 443 | To connect to OpsRamp using its public IP addresses, if the agents are configured to connect directly. |
| TCP port 3128 | To connect to OpsRamp, using the gateway's embedded proxy. |
Agent and gateway connectivity requirements
Recommended hardware requirements
We recommend the following minimum hardware required for Agents:
- 2 Virtual CPUs
- 2 GB RAM
- 200 MB free disk space
- 1 NIC