Working with Interfaces
Every device GridNMS monitors has a list of interfaces — physical ports, virtual interfaces, sub-interfaces, and tunnels. GridNMS discovers them automatically and collects traffic, packet, and error data for each one. But not every interface is worth watching: management loopbacks, unused ports, and chatty virtual interfaces can clutter your charts and inflate your stored data for no benefit.
This page covers how to curate the interface list on a device — showing and hiding interfaces, tidying up the noise, and setting bandwidth thresholds so a saturated link raises an alert before users start complaining.
A device’s Interfaces tab. Every interface the device reports gets a row with its status, speed, VLAN, description and location; select one to open its charts and thresholds.
Finding the interface list
Section titled “Finding the interface list”Open any device from Devices, then select the Interfaces tab on the device detail page. You’ll see a table with one row per interface:
| Column | What it shows |
|---|---|
| Name | The interface name as the device reports it (e.g. GigabitEthernet1/0/1). |
| Status | The port’s current state (for example Up or Down). |
| Speed | The link speed, when the device reports it. |
| VLAN / Voice VLAN | The access VLAN and voice VLAN, when the device reports them or you’ve recorded them. |
| Description | The interface description configured on the device, or one you’ve set yourself. |
| Comment | A free-text note for your team. |
| Location | Optional physical-location note. |
| Chart | A live sparkline of recent traffic. |
Hosts monitored over SSH appear here too. A Linux or Mac host lists the
interfaces it reports about itself, such as eth0 or en0. SSH reports traffic
but not the port state, speed, or VLAN that a switch publishes over SNMP, so
those columns stay empty for a host monitored only that way — see
Metrics explained for what each source
contributes.
Use the search box above the table to filter by name, description, comment, location, or VLAN. This is the fastest way to jump to a specific port on a high-port-count switch. Click a row to see that interface’s full charts and thresholds underneath the list (see below).
Showing and hiding interfaces
Section titled “Showing and hiding interfaces”Hiding an interface removes it from the table and from your charts so you can focus on the ports that matter. Click Manage interfaces next to the search box to show a switch on every row:
- A switch that is on means the interface is visible and being collected — switch it off to hide it.
- A switch that is off means the interface is hidden — switch it on to unhide it.
Once any interface is hidden, the button reads Manage / show hidden (N); click it to reveal the hidden rows (they show dimmed). Click Hide hidden interfaces (N) (or Done) to put the switches away again.
This behavior is deliberate and useful: hiding the dozens of unused access ports on an access switch both declutters the view and trims the amount of data GridNMS stores and processes. The trade-off is that you get no history for a hidden interface, so only hide ports you genuinely don’t need to watch.
When to hide an interface
Section titled “When to hide an interface”- Unused access ports with nothing plugged in.
- Management or loopback interfaces that never carry user traffic.
- Noisy virtual interfaces (for example, large numbers of per-client wireless interfaces) that overwhelm the chart picker.
When to keep an interface visible
Section titled “When to keep an interface visible”- Uplinks and trunks — these are the links you most want to watch for saturation.
- Anything you’ve set a threshold on — a hidden interface collects no data, so its thresholds can never fire.
Broader hide rules (interface ignore overrides)
Section titled “Broader hide rules (interface ignore overrides)”The per-row switch hides one interface at a time. If you want to hide a whole family of interfaces by name — say, every per-client wireless interface — you can use a pattern-based rule instead. Open the device’s Packs tab and scroll to the Interface Ignore Overrides section.
There you can add a name pattern and choose an action:
- ignore — interfaces whose name matches the pattern stop collecting metrics and are hidden (the same effect as the per-row switch, but applied to every matching name now and in the future).
- allow — re-includes a name that a broader ignore rule would otherwise hide. Use this to carve out exceptions.
Rules can be inherited from the device’s class or from its monitoring pack, and a device-level rule always takes precedence over an inherited one. The panel shows you the full picture: what’s inherited, what’s set on the class, and what’s set on this specific device. New rules take effect within about a minute.
Editing interface details
Section titled “Editing interface details”Click an interface row, then Edit details under its charts. This lets you set fields that GridNMS can’t always learn automatically:
| Field | Use it for |
|---|---|
| Description | A friendly description if the device doesn’t supply one. |
| VLAN / Voice VLAN | Record the data and voice VLANs for documentation. |
| Comment | A free-text note for your team (e.g. “patches to rack B, panel 3”). |
Click Save details to apply. These values are yours and persist across re-discovery.
Selecting a row also gives you a time range picker, the full-size Bandwidth chart, and the Thresholds panel described below, plus two smaller charts:
- Packets / s — the inbound (Rx) and outbound (Tx) packet rate. Traffic that looks flat in bits per second but spiky in packets per second usually means many small packets rather than a few large ones.
- Errors / s — the inbound and outbound error rate. This is the one to check when an interface is up and passing traffic but something still feels wrong: a persistent, non-zero error rate points at the physical layer — a marginal cable, optic, or duplex mismatch — rather than at capacity.
Use Pin to dashboard to keep the bandwidth chart for this interface on a dashboard.
Additional addresses on a device
Section titled “Additional addresses on a device”Interfaces themselves are discovered automatically — you don’t add them by hand. But a device often answers on more than one IP address (a management address, a loopback, a secondary subnet). GridNMS lets you record these so the device is recognised no matter which address traffic, syslog, or traps arrive from.
On the device detail page, use Additional Addresses → Add to attach extra IPs to the device. Once added:
- Syslog and SNMP traps arriving from any of those addresses are attributed to this device — not logged as coming from an unknown sender.
- The device is matched on any of its addresses during discovery and lookups.
Per-interface bandwidth thresholds
Section titled “Per-interface bandwidth thresholds”A threshold turns a traffic reading into an alert. When an interface’s inbound or outbound traffic crosses a level you set, GridNMS raises an event — and, if you’ve configured delivery, sends a notification.
To add one:
- On the Interfaces tab, click the interface’s row.
- In the Thresholds panel beside its charts, click Add threshold and fill in the form:
| Setting | What it means |
|---|---|
| Direction | Which traffic to watch — inbound or outbound, as a rate or as a percentage of link speed. |
| Condition | Greater than, at least, less than, or at most. |
| Value | The level to compare against — a rate such as 50Mbps or 1Gbps, or a percentage for the link-speed options. Required. |
| Severity | How serious a breach is, from Critical down to Debug. |
| Custom Message | Optional. Auto-generated from the rule if left blank. |
| Auto-clear | Closes the alert automatically when the value recovers. |
| Active | Turn the threshold on or off without deleting it. |
| Notify on match | Turn on to send the alert to you (Notify me) and/or to one or more notification endpoints — see Notifications. |
- Click Save. Once active, the threshold appears as a reference line on the bandwidth chart, so you can see at a glance how close the interface is running to its limit.
The bandwidth chart has a units toggle above it — switch between Bits per second and % of link speed to see traffic either as a raw rate or as a percentage of the interface’s known link speed. Reference lines follow whichever unit you’re viewing, so a bandwidth threshold shows on the bits-per-second view and a utilization threshold shows on the percentage view. If the interface’s link speed isn’t known, the percentage view shows a note instead of a chart.
Editing and removing thresholds
Section titled “Editing and removing thresholds”The Thresholds panel lists every threshold on the interface. Each row shows its severity, direction, condition and value, a BREACHED marker while it is tripped, and an Inactive marker when it is switched off. Use the pencil and trash icons on each row to edit or delete a threshold. Deleting a threshold removes its alarm immediately; deactivating one keeps the definition but stops it firing.
See Understanding Your Metrics for how thresholds translate readings into alerts, and Events & Alerts for what happens after one fires.
Related pages
Section titled “Related pages”- Understanding Your Metrics — what each reading means and when it signals a problem.
- Overriding Device Details — set your own values for fields GridNMS gathered over SNMP.
- Devices & Inventory — browsing and organizing the device list.
- Events & Alerts — triaging the events that thresholds raise.
docs built 2026-11-12 · 945a2dec2