Skip to content

Nodes

Nodes or Edge Nodes are terms used to refer to the machines that Edge Builder directly manages (where our node components are installed). This section will provide a high level overview of how Nodes are provisioned and managed in Edge Builder. More detailed step-by-step guides can be found in our tutorials section.

Node provisioning

In order to provision a Node in Edge Builder it must first be registered with the Edge Builder Controller (the main Edge Builder server component). Nodes are, in effect, pre-registered or whitelisted. The minimum information required is as follows:

  • Name
  • Description

Once a Node is registered a unique provisioning command is generated. This command is designed to be run on the Node, it will download, install and setup the node components (agent). The Node requires an internet connection and it must be able to address the Edge Builder server components.

While this command can be run manually, Edge Builder offers additional atuomation. If the new Nodes are accesable via ssh, Edge Builder can connect to them and automatically run the provisioning commmand. The following extra information is required to do this:

  • Node IP address
  • Username
  • Password
  • Server IP address

Note

  • While the server IP address is manditory for provisoning via SSH, it can be optinally specified as part of a regular Node provisioning process. This allows you to specify an alternate address for the Node to use to contact the server.
  • One or more labels can be optinally assigned to Nodes during provisioning.

Node management

Edge Builder offers two core features to provide both Node management at scale and full access to individual Nodes where required.

  • Scripts: Edge Builder is able to execute shell or python scripts in parallell across a large number of Nodes.
  • SSH Tunnel: Edge Builder can open an SSH tunnel to any of its managed Nodes. This is reverse ssh tunnel, it does not require any ports to be exposed on the Node.

Node monitoring

Edge Builder monitors the basic status of each Node at all times. Additionally, metrics are capture from the Node. By default these are cached locally (for 48h by default, this is configurable) however, they can forwarded to the server, or set to log live metrics to the server.