This page explains what the stats in portal are showing you, where to find them, how fresh the data is, and where there are known limitations. If a graph looks empty, unexpected, or different to what you'd expect, this page should help explain why.
1. Where to find your stats in portal
Portal stats live in three places:
Home page - two graphs are shown as soon as you log in: LINX Traffic Overview (traffic across all LINX LANs, plus Managed IXP and an Aggregate line) and [Your Organisation] Top 5 Port Services (your five highest-traffic ports). Both link through to a fuller analysis view
My Services - select one or more of your own ports or logical services to see detailed stats for exactly what you run with us. This view has more metrics available than the home page
Peers view (within My Services) - select a port service to see the peers connected to it, with traffic figures per peer, and a combined peering traffic graph
Which view you need depends on what you're trying to find out: the home page is a quick health check across LINX and your top ports, My Services is for looking at a specific port or service in detail, and Peers view is for understanding who you're exchanging traffic with on a given port.
2. Using the controls
Most of the graphs across portal stats share a common set of controls:
Period - choose from Past hour, Past day, Past week, Past month, Past year, or set Custom dates
Metric - switch what's being measured (Bitrate, Errors, Unicast/Multicast/Broadcast Packets, depending on the view, see Section 4 for what's available where)
Time Zone - toggle between Local and UTC
Legend - click an entry (an exchange, port, or peer) to show or hide it on the graph, use the search box above the legend to filter a long list down to what you're looking for
Export - where available, CSV and PNG export buttons sit above the graph, letting you download the underlying data or a static image of the chart
3. What data you're looking at, and how fresh it is
Portal stats are drawn from a few different underlying sources depending on what you're looking at:
SNMP based stats - bitrate, errors, and packet counts, polled at regular intervals from the network devices themselves. This is what powers the LINX Traffic Overview SNMP view, Top 5 Port Services, and the port/service-level stats in My Services
Flow based stats - a separate traffic view (LINX Traffic Overview also has a Flow Traffic tab), built from flow data rather than SNMP polling
Optics (light level) stats - received and transmitted power, collected via a different method and updated approximately every 10 minutes. Coming to My Services in a future release (see Section 7); if you've recently made a change to your port, please allow a few minutes before checking, as it won't reflect the change instantly
Not every graph on the page updates at the same speed. If you've just made a change and a stat hasn't updated yet, it's worth waiting a few minutes rather than assuming something's wrong. We know this isn't always obvious, so we're adding a "last updated" indicator to each stats area, so you can see exactly when the data you're looking at was last refreshed rather than having to guess. This is expected to be live by the end of the year.
For SNMP based graphs, the level of averaging also depends on the date range you select: ranges under a month show 5-minute averages, ranges up to two months show hourly averages, and anything longer shows daily averages. Peak traffic values are always calculated from 5-minute samples regardless of the range selected.
4. Metrics available at each level
Different views expose different metrics:
LINX Traffic Overview (SNMP tab): Bitrate, Errors, Unicast Packets, Multicast Packets, Broadcast Packets
LINX Traffic Overview (Flow tab): Bitrate, Packets. Flow data doesn't distinguish packet type, so this is a single combined count, unlike the SNMP tab, which breaks packets out by type
My Services (your ports and logical services): Bitrate, Errors, Unicast Packets, Multicast Packets, Broadcast Packets
Peers view: Bitrate (shown as combined in/out/total figures per peer)
5. Aggregate view
When viewing more than one port or peer at once, an Aggregate view toggle lets you switch between seeing each one plotted individually, or combined into a single summed line. This is available in both My Services and Peers view. Off (individual lines) is the default.
6. Port groups (LAGs) vs individual ports
If your port is part of a LAG (a group of ports bundled together), selecting the group in portal shows you a combined graph across all the ports in that group.
This means:
A spike, error, or discard shown at the group level doesn't tell you which individual port within the group it came from
If you need to isolate the source of an issue, select the individual port rather than the group
This behaviour is consistent across the ports we support, though the way it's presented may vary slightly depending on the underlying hardware vendor. We'll update this section if we identify vendor specific differences that affect how you should read the data.
7. Peers view
From My Services, selecting a port service and switching to Peers view shows you the peers connected on that port, along with average, current, and max traffic figures (in, out, and combined). Selecting one or more peers builds a combined peering traffic graph below, with the same period, metric, and time zone controls as elsewhere, plus CSV and PNG export.
This is useful for identifying which specific peer is contributing to traffic on a port, rather than just seeing the port's totals.
8. Discards, what they do and don't tell you
Discards are currently only reliably understood on LON1. On our other LANs, discard figures should be treated with more caution, for two reasons:
Background noise: on some LANs, the discard count can include things like proxy ARPs, CPU-uplifted packets, and routing protocol traffic, not just genuinely dropped packets. A non zero discard count doesn't necessarily mean traffic was lost
Hardware limitation: on ports using SR Linux (Broadcom) switching hardware, it's not currently possible to distinguish real discards from control plane discards. There's no fix available for this at present
What this means for you: discard data outside LON1 can still be useful. For example, as a general signal or for spotting anomalies and trends over time, but shouldn't be read as a precise count of dropped packets. If you're trying to diagnose a specific issue, we'd recommend treating a discard reading as a starting point for investigation rather than a definitive answer, particularly outside LON1.
9. Light levels (optics), availability and limitations
The Optics view will show received and transmitted power levels for your port over time. Where a port has multiple lanes assigned (for example, on a breakout configuration), each lane will be shown individually.
Availability: optics data is not currently available on edge core ports (locations: CAR1, JED1, MAN1, SCO1). If your port is one of these, you won't see optics data and this isn't a fault, it's a current limitation of that platform
Update frequency: optics data is collected roughly every 10 minutes, so there will be a short delay between a real world change and it showing up in portal
10. Why a stat might not be showing
If a graph or stat isn't showing what you expect, it's usually one of the following:
Not collected for your hardware - some metrics (e.g. optics on certain edge core platforms, discards accuracy on SR Linux) aren't available due to a hardware or vendor limitation, not a fault
Temporarily unavailable - occasionally the underlying data service is briefly unreachable. If you see a message saying data is temporarily unavailable, this should resolve on its own; try again shortly
Not supported for this port type - some stats only apply to certain port or service types, and won't appear for others
If you're unsure which of these applies to what you're seeing, please contact the NOC and they can clarify.