A NetFlow analyzer that comes with the monitoring, not after it
The interface graph says the uplink is at 94%. It does not say who is doing it. Flow records do, and in SysPeep the collector that reads them is part of the same Windows service that polls the switch.
Key points
- NetFlow v5 and v9, IPFIX and sFlow v5, with listeners on the SysPeep server and collection at remote probes. Sampling rates are honoured.
- A Flows page in the main navigation: exporter status, top talkers, conversations, interfaces, QoS classes, traffic by hour of the week and by country.
- Interfaces are named as the device's SNMP sensors name them, and utilization is shown against the speed the device reports.
- Included in every edition, the free one too. Never priced per interface, per exporter or per flow.
- A router whose export stops raises an incident on that device within minutes, not a quiet gap found later.
On this page: Why flows The Flows page Pointing an exporter at it Honest totals What it costs Where it fits
Most monitoring tools treat flow analysis as the thing you buy next: a module, a per-interface line on the quote, or a second product with its own console and its own login. SysPeep treats it as the second half of the interface counter. The counter tells you the link is busy; the flows tell you which hosts, which conversations and which class of traffic made it busy, in the same console, on the same device page.
Why the counter is not enough
SNMP interface counters are the right first answer. They are cheap, every switch has them, and SysPeep keeps per-interface trend history for the whole estate. But a counter is a total. When the Monday-morning surge or the overnight backup pushes a link toward saturation, the counter can only tell you that it happened. The question the on-call person actually gets asked is "what was it?", and the only honest way to answer is to look at the flow records the router was exporting all along.
What the Flows page shows
The Flows page brings the records from your switches and routers into the console beside the sensors. It shows where to point an exporter, the live status of each one, and the traffic itself, filtered by site, time range and exporter. Heat strips show interface use in and out as a percent of link speed, traffic by exporter and each QoS class's share, one cell per hour over a day or a week and coarser over 30 and 90 days.
A seven-by-24 pattern of the last four weeks shows the Monday surge and the nightly backup window as what they are: a pattern, not a one-off. The busiest senders are placed by country from a bundled database, with no online lookup. Every time on the page is the server's clock, whatever the browser's time zone, so two administrators looking at the same hour see the same hour.
Each monitored device that sends flow carries a flow chip in the device tree and on the map, with "Sending flow" and "Flow stopped" filters, and its device page links each exporting address to its traffic. The flow data is not somewhere else; it is one tab from the device.
On the Flows page
- Exporter status: receiving, no records, not recognised, silent since
- Top talkers, conversations, interfaces, QoS (DSCP) classes
- Heat strips by hour; traffic by hour of the week
- Traffic by country, offline database
- Interface utilization sensors in percent of link speed
- Flow-silence events on the device within minutes
- Unmonitored exporters listed with an Enable button
- Ports, retention and storage ceiling set in the console
Pointing an exporter at it
| Step | What happens in SysPeep |
|---|---|
| Configure NetFlow, IPFIX or sFlow export on the router or switch, aimed at the SysPeep server or at the probe nearest the device | The Flows page says where to point it and which ports the listeners use; the ports are set in the console |
| The first records arrive from a device that is already monitored | Its interfaces are scanned automatically, so "ifIndex 14" becomes the uplink's real name and its utilization is shown against the speed the device reports |
| Records arrive from an address that is not yet a monitored device | The exporter is listed with an Enable button and a daily reminder, not silently dropped |
| You want an alert when the uplink saturates | Two predefined sensors turn interface use in and out into a threshold in percent, seeded with a warning at 80%, with no bandwidth figure typed in per interface |
| The router stops exporting | When every exporting address of the device has been quiet longer than the silence rule, a flow-silence warning event names the silent addresses; an event rule makes it an incident and a notification, and the event closes when flow returns |
| Your branch is behind a firewall | A remote probe (paid editions) collects flows locally and reports back over mutual TLS, so no inbound rule is needed at the branch |
Totals that say what they are made of
Every total on the Flows page says how much of the window the collector actually observed, and an hour the server did not observe is hatched rather than drawn as quiet. This is the same rule SysPeep applies to availability: a gap in the data is shown as a gap, never as good news. Long ranges read hourly totals kept in their own tables with their own retention, so a 90-day traffic view costs the server hourly rows rather than 90 days of per-minute records, and no total is ever drawn from data that was already pruned.
What it is not: a packet capture or a deep-packet-inspection appliance. Flow records describe conversations, not payloads. For the question "which host, talking to which host, on which port, how much" they are exactly the right evidence, and they are cheap enough to keep for months.
What it costs
Nothing beyond the device. SysPeep licenses count monitored devices and leave sensors unlimited, and flow analysis is part of every edition, including the free one. A router you already monitor can export NetFlow from every interface without adding anything to the count. The free edition monitors up to 5 devices at full depth, flows included; the 30-day trial unlocks probes, so flows can be collected at a branch too.
Where it fits
Pair it with the interface counters
The SNMP interface sensors catch the saturation and raise the alert. The Flows page explains it. The bandwidth monitoring page covers the counter side; this page covers the flows. Both live on the same device page.
Learn the subject first
The library's flow monitoring guides explain NetFlow, sFlow and IPFIX from the ground up, and the bandwidth and utilization guides cover how full a link really is and when to upgrade it.
Questions administrators ask about NetFlow
What is the default NetFlow port, and is NetFlow UDP or TCP?
NetFlow, sFlow and IPFIX are all sent over UDP. There is no single default NetFlow port: 2055 is the most common for NetFlow, 6343 for sFlow and 4739 for IPFIX, and SysPeep lets you set the listener ports in the console and offers the firewall rules for them. A NetFlow TCP port is not something you will need to open.
NetFlow v5 vs v9 (templates) vs IPFIX vs sFlow vs J-Flow: which packet format does it accept?
All four. NetFlow v5 is the fixed-format original (IPv4 only); NetFlow v9 is template-based and carries IPv6 and vendor fields; IPFIX is the IETF standard descended from v9; sFlow v5 is sampled rather than complete. Some vendors use their own names for the same thing, such as J-Flow and NetStream, and those are received as the NetFlow or IPFIX versions they really are.
Flexible NetFlow vs traditional NetFlow vs NetFlow-Lite: which should I export?
Traditional NetFlow exports a fixed record; Flexible NetFlow lets you choose the fields, exported as v9 or IPFIX, and is the right choice on devices that support it; NetFlow-Lite is a sampled variant on smaller switches. SysPeep reads the templates the device sends, so any of them works; export v9 or IPFIX when you have the choice.
NetFlow vs packet capture (pcap) and port mirroring: when do I need each?
Flow records summarise conversations: who, to whom, which ports, how much, when. Packet capture through a mirror port records the packets themselves. Flows are cheap enough to keep for months and answer the bandwidth questions; capture is for the one conversation you need to read byte by byte. The packet capture analysis series covers when it is worth the weight.
Configuring NetFlow: the exporter says NetFlow isn't enabled, or no records arrive
On the device, enable flow export on the interfaces that matter, point it at the SysPeep server or the nearest probe on the configured port, and set the active timeout to a minute or less. On the Flows page the exporter shows as receiving, no records, not recognised or silent since, which tells you which side to look at. The flow monitoring series has the per-platform configuration.
Does it collect NetFlow from VMware distributed switches, Open vSwitch, NAS appliances and cloud flow logs?
Any exporter that speaks NetFlow v5 or v9, IPFIX or sFlow is received, which includes VMware distributed switches, Open vSwitch and the NAS and firewall appliances that export flows. Cloud providers export flow logs in their own formats; the cloud flow logs guide covers what they contain and how they differ.
What does the NetFlow data format look like, and can I see a sample?
A record carries the source and destination addresses and ports, the protocol, the interfaces, the byte and packet counts, the start and end times and, in v9 and IPFIX, whatever template fields the exporter chose. On the Flows page you see the records as conversations, top talkers and interface traffic; the raw sample is one click from the exporter's status.
Is there an enterprise NetFlow analyzer price, or a per-interface charge?
No. Flow analysis is part of every edition, including the Free one, and is never priced per interface, per exporter or per flow. A router you already monitor can export from every interface without changing the license count.
Does it support NetFlow for IPv6?
NetFlow v9 and IPFIX carry IPv6 flows; NetFlow v5 cannot, by design, so an IPv6 network needs v9, IPFIX or sFlow export. All are received.
Which is the best NetFlow analyzer software?
For most sites, the one that is already inside the monitoring system that knows the devices, because the question "what is eating the bandwidth" arrives on the same screen that reported the bandwidth being eaten. Standalone analyzers earn their place when flow analysis is a separate team's job.
NBAR2 vs NetFlow: does it identify applications?
Flow records identify conversations by address, port, protocol and QoS class; application recognition by signature (NBAR2 and similar) happens on the device and, where the device exports the application field in v9 or IPFIX, it is received as one more field. SysPeep does not perform deep packet inspection itself.
Is there a NetFlow support matrix: which devices export it?
Most routers, layer-3 switches, firewalls and many NAS appliances export NetFlow, IPFIX or sFlow; access switches often do not. The practical matrix is your own: the Flows page lists every address that is sending records, including the exporters you had forgotten about.
What is in a NetFlow record: the fields, TCP flags, interface index, protocol numbers and file format
The v5 fields are source and destination address and port, next hop, input and output interface index (the SNMP ifIndex, which is why SysPeep names them from the interface scan), packet and byte counts, start and end times, TCP flags, protocol number (6 for TCP, 17 for UDP), type of service and the AS numbers. v9 and IPFIX carry whatever the template defines. There is no file format on the wire; what you keep is the collector's own tables.
Sampled NetFlow vs full NetFlow, flows on a layer 2 interface, and “feature not supported on this platform”
Sampled NetFlow reports one packet in N and the collector scales the totals; full NetFlow reports every flow and costs the device more CPU. Many access switches cannot export flows on a layer 2 interface at all and say so; export from the routed interface or the core, or use sFlow where the switch supports it.
How do I visualize NetFlow data and filter it?
The Flows page is the visualization: top talkers, conversations, interfaces and QoS classes as tables and heat strips, filtered by site, time range and exporter, with every total saying how much of the window was observed. There is no filter syntax to learn; the filters are the page's controls.
Point a router at it and see the top talkers
Download SysPeep, enable NetFlow or sFlow export on a router you already monitor, and the Flows page fills in. Up to 5 devices are free at full depth, flow analysis included; the 30-day trial unlocks every feature except clustering.
Download Free TrialEverything it monitors