Network device monitoring in Saudi Arabia: switches, routers and firewalls in Datadog

When an internal application slows down, the first question is often "is it the network?", and the answer can take an hour of messages between the application and network teams. Network device monitoring is useful when it shortens that conversation.

That needs three things. First, every device that matters is known: core and access switches, routers at branches and data centre edges, firewalls, and the power units behind them. Second, each device reports health and interface load often enough to catch a problem while it is happening. Third, device data sits beside service and server data, under shared tags, so a latency spike can be read against the uplink it uses.

Discovery. The Datadog Agent polls devices over SNMP, either from a list of device addresses or by Autodiscovery, which polls each IP address on a subnet you configure and adds the devices that answer. Datadog documents support for SNMPv1, v2 and v3. We recommend SNMPv3 where devices allow it, because it adds authentication and encryption.

Device profiles. A profile tells the Agent which metrics and tags a device type produces. Datadog ships profiles for networking, security, storage and power equipment in its Supported Devices list. A device with no vendor profile still gets the generic profile, which collects TCP, UDP, IP and interface metrics such as bandwidth use. For a model that is missing, a profile can be extended or written, and Datadog Support can be asked to add it. We check your device models against the list before quoting, so nothing is assumed.

SNMP shows that an interface is nearly full; flow data shows why. With NetFlow monitoring in Datadog, routers, switches and firewalls export flow records to the Agent, which aggregates records that share source, destination, port and protocol into five-minute intervals and sends them on. Datadog then shows traffic by device and interface, top talkers, conversations, protocols and ports.

When SNMP polling is also on, Datadog adds the device name, model and interface names to each flow, which is why we set up SNMP before flows. Flow retention is set per account: Datadog documents 30 days by default, with 15, 60 and 90 days as options.

Datadog notes that one Agent is more likely to drop packets above roughly 50,000 flow packets per second, so on large networks we size the collector hosts and, where needed, split exporters across several Agents.

Most network infrastructure monitoring tools are run by the network team alone, and the application team never opens them. In Datadog, device metrics, traps and syslog arrive in the same platform as servers, containers, traces and logs, under tags such as site, environment and team that you agree once.

A dashboard can then show an application's error rate beside the firewall and uplink it depends on, and an incident review can read device configuration changes against the moment latency rose. For the servers behind those devices, our server monitoring page covers Windows, Linux and VMware hosts. For traffic between services, containers and hosts, which is a different product, see our network monitoring tool page; many estates use both.

Relatedserver monitoring pagenetwork monitoring tool page

• Data center network monitoring. Core and top-of-rack switches, firewalls and the power units in each rack, on one device map, with alerts on interface errors and devices that stop answering.
• Branch networks. Routers and firewalls at branches across the Kingdom, discovered by subnet, with alerts routed to the team that owns each region.
• Hybrid estates. Devices on premises read beside cloud workloads during a migration. Our AWS monitoring page covers the cloud side.
• Replacing scripts and spreadsheets. A hand-kept device list replaced by discovery and monitors that name an owner.

If the one Agent polling your devices stops, monitoring stops with it. Datadog lets you designate an active and a standby Agent, and the standby takes over within 90 seconds if the active one fails.

RelatedDatadog data residency guide

The Datadog cost depends mainly on the number of devices in scope, whether flow data is collected and kept, and how many traps and syslog messages are indexed as logs. Our Datadog pricing guide explains how each product is billed; Interkey quotes the licences with the set-up and support.

The timeline is an estimate, not a promise. A first site, polled over SNMP, often shows data within days once credentials and firewall rules are agreed. A national estate with many branches, flows and custom profiles typically takes a few weeks. The drivers are the number of sites and device models, and how quickly SNMP and port changes are approved.