Documentation

Chat Config

NETALERT includes an optional chat assistant that can answer questions about the data held in the system. It is powered by a language model that runs on a server you provide - the assistant is not a hosted service, and NETALERT does not send data to a third party to use it. Chat Config is where that server is configured.

Navigate to Administration > Advanced Features > Chat Config.

Settings

The page lists the settings as name and value pairs. You edit each one with the button in the Actions column.


Active	Whether the chat assistant is enabled. Set to `0` the assistant is switched off; set to `1` it becomes available in the interface
APIKey	The key used to authenticate against the model server, where one is required. Leave empty for a server that does not require authentication
Host	The URL of the model server endpoint, including the port and path
Model	The identifier of the model to use on that server

Notes

The assistant is disabled by default, and remains so until Active is set to 1 and a reachable Host has been supplied.

Because the model server is one you operate, the Host normally points at a machine inside your own network. This is what keeps the feature consistent with the rest of the product, where traffic capture, analysis and alert generation are all performed on-premises without data being sent to external systems. Pointing Host at an external service would send the content of questions - and the data used to answer them - outside your network, so it should only be done deliberately and with that in mind.

The Model value must be one the configured server actually provides; the name is passed through to that server rather than interpreted by NETALERT. Running a large model requires substantial resources on the host serving it, which is a consideration when choosing one.