Documentation
What NetEnds is
Documents the current release. We keep improving the platform, so some details may change. Ask support if anything looks off.
NetEnds is a network operations and intelligence platform you run on your own infrastructure. It discovers the nodes on your network — switches, routers, firewalls, load balancers and access points — keeps a live picture of their inventory and health, backs up and generates their configuration, and gives you one place to work across many vendors instead of one tool per vendor.
It installs as a set of containers on one Linux host inside your network. Gia, the network assistant that ships with it, runs its model on hardware you control.
The problem it solves
A real network is rarely one vendor. Switches, routers, firewalls, load balancers and wireless all speak different CLIs and APIs, and the tooling usually follows suit: one system for inventory, another for monitoring, another for config backups, another for flow analysis. The picture is spread across all of them, and none of it sits where you make a change.
NetEnds reads each platform through a collector built for it, normalizes what it finds, and presents inventory, monitoring, topology, configuration and logs together. Where a node has no dedicated collector, a standard-SNMP baseline still covers the essentials.
What NetEnds does
Everything below is a destination in the product. Two capability areas belong to a license tier and are marked inline; the rest is the default surface.
-
Discover and inventory. Set Up Network → Guided Setup and Discovery find nodes, identify their vendor, model and OS, and read their interfaces, VLANs, switchports, ARP and MAC tables, routes, VRFs and hardware components. Dedicated collectors cover Arista, Cisco (NX-OS, IOS and IOS-XE, ASA), Juniper, Aruba, Fortinet, Palo Alto and more, plus an SNMP baseline for anything else. Supported platforms has the breakdown by platform and capability.
-
Monitor health and metrics. Observability polls CPU, memory, interface utilization, hardware and power health, routing-protocol state and firewall state, and raises alerts against them. Its Analytics pages add baselines, forecasts and executive summaries.
-
See the network as a map. Topology renders Layer 1, Layer 2 and Layer 3 views, plus multicast and MPLS, so a change has a picture behind it.
-
Back up and change configuration. Config retrieves running configuration on a schedule, versions it and diffs it. It can also generate configuration and push it to a narrower set of platforms. Every push is previewed first and waits for an approver before it runs, and a backup is taken just before it runs as the rollback baseline.
-
Turn on device telemetry. Protocol Enablement renders the device-side configuration that makes a node send syslog, flow records and SNMP traps, and that turns on the management API the collectors read. Autopilot can apply that configuration for you; it stays off until an operator switches it on.
-
Read logs, events and flows. (Events & flows tier.) Logs & Flows ingests syslog and flow records, with search and analytics over both.
-
Ask Gia. (Gia tier.) See below.
-
Automate and test. Automation runs operator-defined playbooks and keeps a record of each run. The Lab designs and deploys emulated topologies so you can try a change before it reaches production; it is part of the Gia tier.
Gia, your network assistant
Gia is built into NetEnds. It answers questions about your network and can take actions, bounded by the permissions of the person asking. You reach Gia in the product, and through a Slack direct-message bridge when the host can reach Slack. It is part of the Gia tier.
Gia’s model runs on your hardware. The product accepts only a loopback or private-network address for the model server, and retrieval against your own runbooks runs on the same server. Apart from the model server contacting its maker, ollama.com, each time it starts, two parts of Gia can reach outside your network, and only when you set them up: the Slack bridge, and knowledge sources in Confluence or SharePoint.
Gia adds about 26 GiB of memory to the host. It runs on CPU, where a multi-step answer takes minutes; with a suitable GPU the same answer takes seconds. It is optional: NetEnds runs fully without it, and you can add it later. System requirements & sizing has the numbers.
Who it is for
NetEnds is for network engineers and operators who run multi-vendor networks and need the platform to live inside their own infrastructure — because of an air-gapped segment, a data-handling rule, or a preference to keep network state off a vendor’s cloud.
Access is role-based. Operators, administrators and auditors each see only the parts of the product their permissions allow, and the license tier decides which capability areas are unlocked.
How it sits on your network
You host NetEnds, and the data it collects is stored in its own databases on your host. It does not depend on a service run by NetEnds, and license checks run entirely on the host.
NetEnds ships integrations that reach beyond your network. Each one does so only after you turn it on and give it credentials:
- Notifications to Slack, Microsoft Teams, email, SMS, WhatsApp and generic webhooks.
- Tickets to Jira, ServiceNow and generic webhooks.
- Gia’s Slack bridge and its Confluence and SharePoint knowledge sources.
- External sign-in through the LDAP, RADIUS, SAML or OIDC service you name.
New releases reach you as files from NetEnds, the same way the first install does. Updating NetEnds covers how to apply them.
Where to go next
- Supported platforms — which platforms NetEnds can discover, monitor, back up, configure, upgrade and enable, and how.
- System requirements & sizing — the host to plan for.
- Installing NetEnds — from the release email to the first sign-in.