Network monitoring utilities, run as a managed service
Network monitoring is the continuous checking of switches, routers, firewalls, access points, servers and links, so faults are found before users report them. Arke Technologies, in Johannesburg, selects and runs network monitoring utilities for organisations across South Africa, turns alerts into tickets and reports to you every month under an SLA.
What is network monitoring?
Network monitoring is the practice of watching a network's devices and links continuously, measuring their health and performance, and alerting the right person when something fails or drifts outside agreed limits. IT network monitoring answers three questions at any moment: is it up, is it fast enough, and has anything changed that should not have?
Network monitoring definitions vary between vendors, but a complete service covers availability (can each device and link be reached), performance (latency, jitter, packet loss and utilisation), traffic (who and what uses the bandwidth), configuration (what changed, and who changed it) and security events. Monitoring network performance over months also shows when a link or switch needs upgrading, before it becomes a complaint.
Network monitoring and management go together: monitoring finds the fault; management (changes, firmware and capacity planning) fixes it and stops it recurring. Monitoring is part of our wider network infrastructure service.
Describe the key features in network monitoring tools
Asked to describe the key features in network monitoring, we look for these, whatever the product:
Discovery and topology maps. Devices found automatically, with a map of how they connect.- SNMP polling. Counters read from switches, routers, firewalls and access points, using SNMPv3 because older versions send data in clear text.
- Flow analysis. Monitoring network traffic with NetFlow, sFlow or IPFIX shows which applications and users consume bandwidth.
- Log collection. Syslog from network devices, kept centrally with synchronised time stamps.
- Dependency-aware alerting. When a site router fails, you get one alert, not fifty.
- Dashboards. Real time network monitoring views; many teams build network monitoring with Grafana on top of the collected data.
- Scheduled reports. Availability, capacity and trends.
- Integration. An API or webhooks, so alerts open tickets in a service desk automatically.
- Access control. Named accounts and roles, so the monitoring tool is not itself a weak point.
A network management system (NMS) combines most of these. A single network monitoring utility, such as ping or traceroute, does one job well and nothing else.
NOC, SOC and what we do not claim
A network operations centre (NOC) keeps networks and services available; a security operations centre (SOC) detects and contains attacks. NOC network operations are about uptime and performance, while SOC work is about intruders. Many providers sell NOC and SOC services together, and a cyber security NOC is usually a NOC that also reads firewall and intrusion alerts.
We are specific about what we provide. Our monitoring is automated and continuous; human response follows the hours in your SLA, with out-of-hours cover by agreement. We do not staff a round-the-clock operations centre, and we do not claim to. For security network monitoring we collect firewall, VPN and sign-in logs, and for clients who need a SIEM we deploy Microsoft Sentinel; see managed detection and response and firewalls and network security.
AI for network monitoring
AI for network monitoring is useful for two jobs: spotting unusual patterns in traffic or latency that fixed thresholds miss, and grouping related alerts into one incident. Where we use it, it runs under the same rule as all our AI work: AI recommends; people decide. Every AI-suggested action is logged and reviewed by an engineer before anything changes on your network.
Network monitoring checklist (use it as a template)
These network monitoring best practices form our onboarding checklist; use it as a template for your own review:
- Keep a complete inventory of network devices, with owners and locations.
- Use SNMPv3 and named accounts, and retire default community strings.
- Synchronise every device's clock (NTP), so logs line up.
- Back up device configurations automatically, and after every change.
- Set thresholds from your own baseline, not vendor defaults.
- Configure dependencies, so one outage raises one alert.
- Schedule maintenance windows, so planned work does not raise false alarms.
- Keep an escalation list with names, not only roles.
- Set log retention to meet POPIA and your audit requirements.
- Review the report every month and act on the trends.
Reporting, governance and compliance
- The network monitoring report. Monthly: availability by site and device, incidents and their causes, capacity and utilisation trends, changes made, open risks and recommendations.
- Service governance. An ITIL-aligned service desk practice: an SLA with response and resolution times, a named service lead, escalation paths, change control, documented runbooks and a monthly service review.
- Compliance. Monitoring logs can hold personal information such as user names, device names and IP addresses, so we handle them under POPIA with defined retention and access. Logging is aligned with ISO/IEC 27001 and the CIS Controls, supports King IV Principle 12 reporting to boards, and gives public-sector clients evidence for PFMA and MFMA audits.
- Procurement. Arke Technologies (Pty) Ltd is a Level 1 B-BBEE contributor (sworn affidavit available on request); company documents, our CSD supplier number and SARS tax compliance status are available on request. See our company facts.
- Continuity and exit. The monitoring configuration, history and reports are yours; when a contract ends we export them and hand over administrator access.
What network monitoring costs
Network monitoring is usually part of a managed IT services agreement priced per user per month, set after an assessment; a network-only SLA is quoted separately, after an assessment of your organisation's needs. What drives the cost:
- the number of devices, sites and links monitored;
- tool licensing, since open-source platforms carry no licence fee while commercial tools are often licensed per sensor or device;
- log retention and reporting requirements;
- response hours, including any out-of-hours cover;
- whether fault fixing is included or quoted per project.
For context, published South African managed IT rates run from about R180 to R400 per user per month, and IT project work from about R550 to R750 an hour ex VAT (matnet.co.za, September 2026).
Last reviewed: September 2026.