What is Network Monitoring?

What is network monitoring?

Network monitoring is the continuous observation of the network that connects your servers, services and users: devices such as routers, switches and firewalls, the links between them, and the traffic crossing those links. Its purpose is to spot a saturated link, a packet-dropping interface or a failing route before users notice slow pages or failed API calls.

The output is typically a topology view, per-interface graphs, alerts on thresholds such as bandwidth use or error counts, and flow records (NetFlow, sFlow, IPFIX) that answer "what traffic crossed this link at 14:32". Tools in this space include Zabbix, LibreNMS, Nagios and PRTG.

Network vs infrastructure vs application monitoring

The three layers overlap in tooling but answer different questions:

• Network monitoring watches the pipes: link use, packet loss, routing changes, DNS resolution. It answers "is the network slow?"
• Infrastructure monitoring watches the hosts: CPU, memory, disk, containers. It answers "is this server overloaded?"
• Application performance monitoring watches the code: per-endpoint latency, slow queries, error rates. It answers "why is checkout slow for this customer?"

A complete operations setup uses all three.

What network monitoring covers

• Availability: ping (ICMP), SNMP up/down state and port reachability checks at fixed intervals.
• Bandwidth and throughput: bytes per second per interface, peak against average, percentile use over time.
• Loss, latency and jitter: measured with active probes or passive flow analysis; the three things that hurt voice, video and real-time traffic first.
• Routing and protocol health: BGP session state, OSPF neighbour changes.
• Flow data: per-conversation records for capacity planning and forensics.
• Configuration drift: device configs saved and compared, with an alert on unexpected changes.

Key network monitoring metrics

1. Availability: percent of time a device or link is reachable.
2. Bandwidth utilization: share of link capacity in use, often read at the 95th percentile.
3. Packet loss: dropped packets as a share of packets sent.
4. Latency and jitter: round-trip time and how much it varies (see latency).
5. Error and discard counters: CRC errors and output drops; sustained non-zero values point to cabling, optics or buffering trouble.
6. DNS resolution time: a slow resolver delays every first request to a host.

Example and how to run it

Users in one region report a slow checkout, while the application dashboards look normal. The network view shows the uplink for that region at full capacity with output drops on one interface. The cause is traffic, not code, and the fix is capacity or traffic shaping.

To set it up: poll devices with SNMP, or streaming telemetry on newer gear, and store the time series. Add active probes between sites for latency and loss, and a flow collector for per-conversation visibility. Build dashboards by site and service, and alert on what maps to user impact (loss, latency, saturation) instead of every counter.

How it relates to load testing

When a load test shows rising response times, network monitoring helps tell whether the limit is in the network, such as a saturated link or packet loss, or above it, in the application or database. Read both signals together before a launch. The guides on network jitter in cloud load testing and network protocols in performance testing go deeper.

What LoadFocus does and does not cover

LoadFocus is not a network device monitor: it does not poll routers or switches over SNMP and it does not collect flow data. What it does is check from outside, as a user would. API monitoring runs scheduled checks from a choice of 25 cloud regions and splits each response time into wait, DNS, TCP connect, first byte and download, and it also offers TCP checks (is a port open and accepting connections) and DNS checks (do your records resolve to the right values). Free checks run every 5 minutes. This is a form of synthetic monitoring, and it shows whether users in a given region see a problem, not which device causes it.

FAQ: Network monitoring

What is network monitoring in simple terms?

It is watching the health of the network that connects your servers and users, so you can find a failing link or device early. It measures availability, bandwidth, packet loss and latency.

What is the difference between network monitoring and network management?

Monitoring observes and alerts. Management includes changing the network: configuring devices, applying updates and fixing faults. Many tools combine both.

Which protocols does network monitoring use?

Common ones are SNMP for device status and counters, ICMP for ping, NetFlow, sFlow or IPFIX for traffic flows, and streaming telemetry on newer equipment.

Does network monitoring replace application monitoring?

No. It shows whether the network is healthy, not why a specific endpoint or query is slow. You need infrastructure and application monitoring as well.

Can LoadFocus monitor my routers and switches?

No. LoadFocus checks websites and APIs from the outside, with HTTP, TCP and DNS checks from cloud regions. It does not use SNMP or collect flow data.