Agents
The agent is a software package that runs on a system and at a high level does the below.
- Control: Starts, stops and monitors programs
- Communication: Sends and receives data and commands from the cloud
- Access: Optionally provides remote access to the system
INFO
Only one agent can run on one system at one time.
INFO
In hard real time scenarios the agent is responsible for ensuring the time constraints are met by programs.
Core and capabilities
From version 1.0.24, the agent is built as a hardened Core runtime with optional capabilities layered on top.
Core is always present and covers program deployment, logging, device control, and Horizon Zero Trust. Remote access features are optional and are chosen when you install the agent:
- CloudLink: Encrypted WireGuard tunnel for remote network access to the device
- Remote terminal: Interactive browser shell that runs as root on the device
A Core install carries neither. Capabilities are fixed when the agent is built, so changing them means reinstalling the agent.
For more information, see Capabilities.
How is the agent managed?
The agent is managed in one of two ways.
- Mutexer Cloud Portal: Through the online Agents window.
- Via CLI: Via command line in the terminal of the device.
Typical actions may be to install, uninstall and upgrade the agent.
How will my system be affected?
When installing and running an agent it is important to understand the changes to the system.
The agent:
- Requires root privilege
- Starts up on boot
- Logs out to syslog
- Modifies iptables
- Installs docker
- Installs wireguard, only when the CloudLink capability is included