Network monitoring
Also known as: NPM, NetFlow, SNMP monitoring, LibreNMS, Zabbix, PRTG
Network monitoring tools continuously watch routers, switches, servers, and links, collecting metrics like bandwidth utilization, error rates, latency, and availability.
When something goes wrong (a link goes down, a device becomes unreachable, a threshold is crossed), alerts fire for immediate response.
Network monitoring, explained at three levels
At home
Network monitoring is like a smoke alarm for your network. Instead of waiting for users to report problems, monitoring tools continuously check whether devices are reachable, links are running, and traffic is flowing normally, alerting IT when something looks wrong before it becomes a bigger incident.
At work
IT operations teams use network monitoring to detect and respond to issues proactively. A monitoring alert saying 'internet link utilization at 95%' lets them address capacity before users experience slowdowns. Without monitoring, they find out about problems from user complaints.
At home
Network monitoring typically covers: availability (is this device/link up?), performance (bandwidth utilization, error rates, latency), configuration (has anything changed?), and capacity (are we approaching limits?). Common tools: LibreNMS, PRTG, SolarWinds NPM, Nagios, Zabbix. Most use SNMP for device metrics, ICMP ping for availability, and flow data (NetFlow, sFlow) for traffic analysis.
At work
NetFlow/IPFIX export from routers provides granular traffic analysis, which IP addresses and ports are consuming bandwidth, traffic matrix between sites, and security anomaly detection. Packet capture (Wireshark, tcpdump) provides deep inspection for troubleshooting. Network performance management (NPM) tools create baseline behaviour profiles and alert on deviations rather than just static thresholds.
At home
Streaming telemetry (using gRPC and OpenConfig models) is replacing SNMP polling for high-frequency metrics from modern routers (Cisco IOS-XR, Juniper Junos). Telemetry pushes data from devices at configurable intervals (as frequent as 1 second) rather than waiting to be polled, enabling faster anomaly detection. Assurance platforms (Cisco Catalyst Center, formerly DNA Center, and Juniper Apstra) add intent-based verification: comparing actual network state to desired state and flagging deviations.
At work
AIOps for network monitoring uses machine learning to correlate events across devices (a router CPU spike + interface errors + latency increase may indicate an imminent hardware failure), reduce alert storms (correlating thousands of events to a single root cause), and predict failures before they occur. Topology-aware monitoring understands device dependencies: if a core switch fails, its downstream failures are suppressed from alerting since they're a consequence, not an independent event.
Real-world examples
At home
Marcus runs LibreNMS on a Raspberry Pi at home. It monitors his router, NAS, and smart home hub via SNMP. He gets a text message if any device goes offline: useful for catching when the NAS has a problem before he loses data.
At work
Nadia's network team uses SolarWinds NPM to monitor 500+ devices. Weekly capacity reports flag links approaching 70% average utilization for upgrade planning. Alerts reach the on-call team within 60 seconds of an outage, with automated ticket creation.
Common confusions
- Network monitoring and application performance monitoring (APM) are the same — Network monitoring watches infrastructure: switches, routers, links, bandwidth. APM watches application behaviour: response times, error rates, distributed traces. They complement each other: network monitoring finds infrastructure causes; APM finds application-layer causes.
- Monitoring alone prevents outages — Monitoring detects and alerts: it doesn't prevent failures. Good monitoring reduces MTTR (Mean Time to Repair) by accelerating detection and diagnosis. Preventing outages requires redundancy, change management, and capacity planning. Monitoring tells you what's happening; the team must act on it.
For analysts
Network availability data from monitoring systems feeds SLA compliance reporting and vendor accountability reviews. BAs can use monitoring data to calculate actual uptime vs. contracted uptime and quantify the business impact of connectivity outages.
History & origin
Network monitoring grew with enterprise networks in the 1980s and 1990s. SNMP (1988) became the standard data collection protocol. HP OpenView and IBM NetView were early enterprise tools. Open-source tools like Nagios (1999) and Zabbix democratized monitoring. The rise of cloud infrastructure and DevOps shifted monitoring toward time-series metrics and observability platforms (Prometheus, Grafana, Datadog) complementing traditional SNMP-based network monitoring tools like LibreNMS and PRTG.