Table of Contents
Why monitor your network?
Businesses rely on accurate network monitoring data because the network is the backbone of your IT infrastructure. Lacking the means to communicate internally or externally about your network can be a disastrous situation, especially if you provide digital goods or services.
Network monitoring tools shouldn’t be a “nice to have” thing which may or may not make this year’s department budget. Network monitoring tools are essential in order to keep tabs on performance, spot anomalies and identify potential security issues. Monitoring empowers your engineers to catch complications before they develop into problems and help preserve high availability.
Network monitoring tools can also help build a faster, more reliable network: once network traffic is characterized by baselining and collecting data, you can reasonably predict the effect of additional services or computers on the network, allowing the chance to build in low latency and keep to Quality of Service metrics for vital systems and applications. If you’re not monitoring your network, you’re not managing it.
What network monitoring tools are available?
Today’s network monitoring tools are diverse and can be generally grouped into two groups:
- free, open source solutions
- specialist turnkey solutions
In a Microsoft ecosystem, you already have the SNMP application-layer protocol in place which can be enabled on network devices to collect and organize information. On the server side, Systems Centre Operations Manager can also be used to monitor networks and servers. Microsoft’s Network Monitor is a free legacy tool that can be used to troubleshoot network problems.
The Linux ecosystem offers a large selection of network monitoring tools such as Monitorix, Darkstat, nload and Netdisco.
Common open source software used for monitoring include Graphite, which can be configured to run on clients or can use common network protocols for monitoring and Nagios, while SmokePing and MRTG are free network data tools and WireShark is a popular, free packet analysis tool.
Your installed network equipment should also have manufacturer tools and software available. For instance, WatchGuard’s Firebox firewalls include network monitoring and analysis tools, Ubiquiti has the UISP network management system and Palo Alto firewalls use the Application Command Center dashboard to provide intelligence about network traffic.
Some well-known specialist network monitoring solutions include PRTG and Solarwinds. PRTG is agentless software which can monitor and classify system conditions like bandwidth usage or uptime and collect statistics from miscellaneous devices and applications. SolarWinds spins out the various moving parts of a network to monitor performance, traffic analysis and configuration.
For cloud-enabled infrastructure, Graphite and Prometheus are free, open-source monitoring solutions that come with their own visualization capability.
Graphite is an open-source time-series data monitoring tool with interactive dashboards and reporting capabilities. Time-series data is a sequence of numerical data points, information snapshots that can be stored and retrieved.
Graphite can render time-series data on demand and is a scalable monitoring tool designed to be deployed and run on inexpensive hardware. Graphite’s integrated web app allows the generation of graphs from retrieved data and also provides a user interface to help navigate various types of metrics and build user-defined dashboards.
Graphite is ideal for cloud-enabled technology teams: Ops and DevOps engineers can use Graphite to continuously monitor their applications and servers.
Hosted Graphite, part of MetricFire, specializes in Graphite monitoring and its product can be used with minimal configuration to gain in-depth insight into your IT environment. Although Graphite is open source software, installing and maintaining it takes time and energy. Hosted Graphite can host your data for you, taking the cost and time-sink of installing and configuring storage equipment and servers off your hands.
How should I set up my network monitoring?
Network monitoring can be split into real-time monitoring and alerting, and gathering data for trend analysis.
Some network monitoring solutions use agents - software installed on monitored devices that reports information. Agents can potentially have access to more detailed information and process it locally, reducing server burden. Although lightweight, some agents may also impact performance on a device. Other monitoring agents do away with agents, preferring to pull data centrally. Agentless solutions are easy to set up and maintain as there’s no agent installation or management required.
Before putting a monitoring solution in place you should establish a baseline reading of network traffic which can be used to spot anomalies later on. Without a baseline, monitoring falls short because you have nothing to compare data with. Knowing the baseline will also help with future upgrades and assessing the impact of additional strain on the network.
Your monitoring system should tell you when the network goes down and when it's back up again and should also inform you of network routing issues. Real-time alerts allow you to quickly react to problems, potentially saving your company money by avoiding downtime. You should consider setting up alerts based on unusual conditions - such as spikes or traffic drops - that might indicate a problem. These alerts can also be configured to specific parameters for your environment in order to reduce alert noise and they can be targeted so that technical teams receive alerts applicable to them and their designated systems.
Gathering historical network data from specific metrics can be used to build statistics for management or customers and highlight behavioral anomalies that can indicate problems. Common network performance metrics can include packet loss, availability, connectivity and throughput.
Good network monitoring tools should also include features that compare data and spot trends that will give you a heads-up on potential network problems as well as getting to the cause faster.
Data gathered from monitoring can also help spot anomalies in network traffic, acting as threat detection and for troubleshooting.
Ideally, your network monitoring solution will allow you to monitor everything on one easy to read visual dashboard across all locations and even across different vendors. With a single view, there’s no need to log in to different platforms to monitor various issues, saving time.
Network monitoring can also aid in network optimization: once you have a clearer picture of network utilization, you can eliminate superfluous services to free up bandwidth for core functions.
Networking monitoring is an important part of network management. Using great network monitoring tools has the potential to save time and money by giving a heads-up on potential network outages and spotting possible security breaches before they happen. Your IT team can keep a close eye on performance with tailored alerts and pull hard data from network monitoring systems to identify capacity trends and assist with future upgrades.
When set up correctly, monitoring can help proactively discover underlying issues to prevent downtime incidents or even improve incident response times, dramatically improving outcomes for better performance and business profitability.
When choosing a solution to evaluate and eventually implement, you should take into account budget, convenience, scope and scalability. Feel free to explore solutions in the cloud or from different vendors if you have been tied in to a legacy vendor for too long.
Numerous network monitoring solutions are available, from open-source configure-it-yourself software to tailored turnkey solutions that meet your environment’s specific parameters, whether on-premise or in the cloud. MetricFire brings the best of both worlds to the table by offering a hosted open-source solution.
MetricFire provides a Hosted version of Graphite which includes storing your data for two years, the Grafana data visualization tool, and much more. Minimal configuration is required to gain in-depth insight into your environment. If you would like to learn more, you can easily book a demo or sign up for the free trial.