What is server monitoring?

A web server is the basis of every web project: the machine where the files, the application and the data live. Users send requests to it and get responses back, so a server that is unreachable is a project that is unreachable. Left unchecked, a server eventually goes down for one of a long list of reasons, which is why monitoring belongs to normal maintenance rather than to incident response.

What is server monitoring

Server monitoring means checking availability, performance and resource use continuously, so problems are caught before users notice them rather than found afterward in a log. It is part of maintenance, not an optional extra. Monitoring the request and response sequence of an Apache server, for example, tells you whether the server software itself is healthy. A complete setup also covers server uptime monitoring, speed, accessibility and the database behind the application.

A well chosen platform handles all of that, lets you react quickly to a live issue, and keeps a history you can read for the slow problems that only show up over weeks. There are many monitoring products, and they differ in what they can see, so the choice matters.

What does server monitoring check

Server monitoring splits roughly into two layers. The first is availability and response performance: is the server reachable, how fast does it answer, is the service behind a given port working. The second is server performance in the narrower sense, the hardware and OS-level resources every service on the machine shares, namely CPU load, RAM usage, disk space and I/O, and network throughput. A server can be technically up and still be minutes away from an outage if one of those resources is maxed out, which is why both layers matter together and not uptime on its own. HostTracker's CPU, RAM and disk monitoring covers the resource layer alongside the availability checks.

Server monitoring tools are software that runs those checks, varying in which metrics they cover. They test accessibility, stability, software functionality and network integrity, and most present the results through an application or a web dashboard.

Most of them work externally, sending requests of the type that suits the metric. Ping, the most basic reachability tool, sends an ICMP echo request and waits for the matching echo response.

A server monitoring service is a set of such tools working together on performance, accessibility, stability and the rest. Because the checks are combined, a service gives deeper diagnostics than a single tool run by hand, runs them on a schedule, and the better ones notify whoever maintains the server when a check fails.

What types of servers can be monitored

There is no real restriction on server type. Not every metric is available on every machine, but you can check the performance of a database, software, SaaS or mobile server, and the same holds for how the resources are allocated: shared, virtual, dedicated. If it has an IP and answers from the network, you can monitor it.

Agent-based versus agentless monitoring

Monitoring tools fall into two families. Agent-based monitoring installs a small program on the server itself, which reads internal metrics such as CPU, RAM and disk usage on the machine and reports them out. Agentless monitoring tests the server from outside, the way a real visitor or client would, sending an HTTP request, a ping, a database query or a port connection and measuring what comes back, with nothing to install or keep updated on the server.

Availability, response time and functional checks (is the site up, does the login form work, does the database respond) are naturally agentless, since the point is to see what an outside user sees. HostTracker runs these checks from a network of checkpoints that reach the server from outside, which is also how it catches problems that appear only from certain regions or networks. An agent sitting on the server itself never sees those.

Software server monitoring

A software server is hardware running specific software to support other servers or computers on the same network. Web hosting and mail servers are typical examples. They need no special treatment, so any service that covers general server monitoring will report on their status and metrics.

SaaS server monitoring

A SaaS server is a cloud server built for one purpose, giving access to an application from anywhere on the network. Monitoring it is close to identical to monitoring any cloud server, with no separate routines to learn, so anything that handles cloud servers handles this too.

Mobile server monitoring

Mobile servers run on smartphones through purpose-built software, using a conventional hardware server as a gateway while the server itself lives on the device. The usual external methods do not apply. You monitor them with built-in plugins or with services written specifically for mobile servers, and the latter are rare.

What to alert on

A monitoring setup is only as useful as the alerts it triggers, and alerting on everything trains everyone to ignore notifications. A practical baseline to alert on includes:

• The server or site going unreachable, or a check failing from multiple checkpoints at once
• Response time crossing a threshold that matters for your users, not just a full timeout
• CPU, RAM or disk usage sustained above a set percentage for longer than a few minutes
• An unexpected drop in successful requests, or a spike in error responses
• SSL certificate or domain expiration approaching

Escalation matters as much as the threshold. A first alert to the on-call contact, escalating to the rest of the team if nobody acknowledges it, is what keeps a real incident from sitting unnoticed. See HostTracker's full feature set for how monitor types, checkpoints and alert channels combine into that kind of setup.

How to do server monitoring

The first step is picking a service that covers every check you need, since the products differ widely in scope. HostTracker is one of the broader ones, with 13 monitor types covering performance, network and availability.

Getting started means registering and configuring the checks you want, from performance and network checks to server uptime monitoring. From there you set the check intervals, the notification schedule and the rest of the quality-of-service settings.

The best server performance monitoring service

HostTracker has monitored websites since 2004 and now watches over 500,000 websites. It starts with a free trial, so you can test the checks against your own servers before paying for anything.

Alerts reach whoever is responsible for the server at the moment the issue occurs, and you choose which users get notified. Support runs around the clock. For the resource layer specifically, see the CPU, RAM and disk check guide, and the rest of the monitoring concepts section for the terms this guide builds on.