Reading the Pulse of elementary OS with Its Built-in System Monitor
When an elementary OS installation starts to feel sluggish, the first instinct is often to blame the distribution itself, swap window managers, or reinstall from scratch. In most cases, however, the answer is sitting quietly in the Applications menu: a small, unassuming utility that has been part of the Pantheon desktop since the earliest Loki builds. The System Monitor, sometimes called the elementary Task Manager by people fresh from Windows or macOS, is a GTK application that lets you see exactly what the machine is doing at any given moment. Learning to read it well is the difference between guessing at the cause of a slowdown and actually knowing whether the bottleneck is the processor, the memory, the storage, or the network.
For readers who are running elementary OS on hardware bought locally from a Sydney retailer or a Melbourne refurbisher, this matters even more. Australian notebooks, particularly the popular mid-range ASUS, Lenovo, and Acer models sold through JB Hi-Fi and Officeworks, often ship with thermal envelopes that react strongly to ambient temperature. A computer in a Brisbane home office in February can throttle for reasons that have nothing to do with software, and the only way to tell the difference is to look at the numbers.
Opening the Tool and Understanding Its Three Tabs
System Monitor lives in the Applications launcher under System, with a small icon that resembles a stylised heartbeat. Clicking it brings up a single window with three tabs across the top: Processes, Resources, and File Systems. Each tab answers a different question. The Processes tab is the equivalent of opening the bonnet of a car to see which cylinders are firing. The Resources tab graphs the activity of the CPU, memory, and network over time. The File Systems tab lists every mounted drive, its total capacity, used space, and remaining space.
The visual language is restrained, in keeping with the broader elementary aesthetic. There are no flashing red bars, no warning sirens, and no pop-up prompts begging for an upgrade. Numbers update roughly once every two seconds, which is fast enough to catch a runaway process but slow enough that the columns can actually be read. Right-clicking a row in the Processes tab offers three options: stop the process, kill it, or change its scheduling priority. For users working between Adelaide and the eastern states, where network latency to common cloud services can vary, this responsive layout means the tool rarely needs the cursor to be precisely placed.
A small detail that often goes unnoticed: each column header in the Processes tab is clickable and will sort the table. Tapping on CPU, for instance, instantly reorders the list to put the heaviest consumer at the top. This is a far more efficient way to find a culprit than the old ritual of opening a terminal and running top or htop, especially on a tablet or two-in-one device. It also helps when the system is being driven by a less experienced user, because the same result can be obtained with a single click rather than a remembered incantation.
Reading the Resources Graph for Real Bottlenecks
The Resources tab is where most diagnostic work actually happens. Three coloured lines move across the window: a blue line for processor activity, a green line for memory, and an orange line for network throughput. When the machine feels slow and all three lines are flat, the problem is almost certainly not elementary OS itself. It might be a slow drive, a failing cable, or a service that the user does not know is running. When the blue line is pinned near the top, the answer is usually a single process that is hogging the cores.
A common pattern in Australian homes is the household where several devices share a connection through a router that links to the National Broadband Network. The NBN, whether delivered as fibre to the premises, fibre to the curb, or the Sky Muster satellite service for remote properties in the Kimberley or along the Nullarbor, has its own latency and throughput characteristics. If the orange network line on System Monitor spikes every time someone in the house starts a video call, the user can see at a glance that the slowdown is bandwidth-related rather than a local issue. This is a more honest diagnosis than the usual complaint that surfaces in community forums after a software update.
Memory is the trickiest of the three. elementary OS uses a fairly aggressive disk cache, which means a healthy machine can show a green memory line that looks worryingly high, often above 80 per cent. The figure alone is not a problem. What matters is whether the system is also swapping heavily. By switching to the File Systems tab and watching the activity column on the root partition, a user can quickly confirm whether the disk is being thrashed by swap. If it is, the memory pressure is real. If it is not, the cached memory is simply doing its job and can be safely ignored.
Identifying the Process That Is Actually Misbehaving
Once a bottleneck is suspected, the Processes tab is the next stop. Three columns deserve attention: the process name, the percentage of CPU it consumes, and the amount of memory it holds. Most users will recognise familiar names like Web, the Pantheon file manager Files, or the Mail client. Anything else, particularly long alphanumeric strings or names with no clear owner, is worth investigating. A simple web search for the process name usually reveals whether it is benign or part of something that should not be there.
It is worth pausing here on the kill function. The Stop and Kill options in the right-click menu are not interchangeable. Stop asks the process politely to terminate, which gives it a chance to save its state, close files, and shut down cleanly. Kill sends an immediate signal that the kernel interprets as a hard stop, useful only when a process has frozen and is ignoring the polite request. Killing a process that is writing to a database or a network share can corrupt data, so the choice matters. The same logic applies on a workstation in a small Perth accounting firm as on a home laptop in Geelong: respect the data first.
For processes that repeatedly appear at the top of the list, a useful follow-up is to open a terminal and run ps aux | grep processname to see when it was started and what arguments it was launched with. elementary OS, like other Debian-based distributions, stores startup services in systemd units, and disabling a problematic one is often a matter of running sudo systemctl disable --now name.service. System Monitor, however, will not show this step. It will only show the symptom, which is why combining it with a quick command-line check is often the fastest path to a fix on a stubborn machine.
Watching Temperature, Power, and the Local Climate
Australia has some of the highest retail electricity prices among developed economies, and the combination of summer heat in cities like Melbourne, Sydney, and Darwin with rising air-conditioning use means that a laptop left on a desk in the afternoon sun can climb several degrees above its comfort zone. elementary OS does not, by default, expose CPU temperature in System Monitor. To see this, users can install a GNOME extension that adds temperature sensors, or run the sensors command from a terminal after installing the lm-sensors package. The graph in the Resources tab does, however, reveal a thermal-throttling pattern: a CPU line that sits high, then suddenly drops, then climbs again, is often a chip protecting itself from heat rather than a software fault.
This is a good moment to consider the broader cost of background processes. A process that uses 5 per cent of a CPU continuously is, in human terms, a small heater running all day. Over a year, on a machine that is never switched off, the difference between a clean system and a polluted one can show up on the quarterly bill from AGL, EnergyAustralia, or Origin. Watching the Resources tab while idling, with no applications open, gives a baseline. Anything consuming a noticeable slice of CPU at idle is a candidate for removal or replacement, and over a few years that adds up to a tangible saving on a household energy budget.
Storage wear is another Australian-relevant concern, especially for users who bought a solid-state drive locally and want it to last the standard five to seven years. The File Systems tab shows read and write activity on each partition. Sudden spikes that are not explained by a copy operation often point to a logging daemon, an indexing service, or a backup tool that is overdoing its job. elementary OS does not ship with a particularly heavy indexer, so any unusual disk activity is usually easy to track down through the Processes tab and resolved without much fuss.
Using System Monitor as a Daily Companion
The real value of System Monitor is not as a single-use tool but as a regular companion. Opening it once a day for thirty seconds, particularly after installing a new application or after a system update, builds a mental baseline of what the machine should look like when it is healthy. When something does go wrong, the user can compare the current view against that baseline and immediately see what has changed. This is the approach that most experienced users in the elementary community eventually adopt, and it is the same approach that the maintainers themselves describe in forum posts when they help newcomers triage mysterious slowdowns.
A practical habit is to leave System Monitor pinned to the dock with the option to show resource usage as a small graph. The right-click context menu on the dock icon exposes a compact version of the Resources tab, which can be set to display CPU, memory, or network usage as a moving bar. For a developer in Surry Hills working from home, or a student at the Australian National University in Canberra, that single bar is often enough to know whether a build is still running or whether a process has stalled mid-compile. It is a small visual cue, but it is the kind of low-friction feedback that turns a debugging session from a chore into a quick check.
Over time, the same three numbers become a kind of vital signs chart for the machine. A sudden jump in idle CPU, an unexplained drop in available memory, or a disk that fills up faster than expected are all signals that something has changed, and the System Monitor is the first place to confirm them. Once a baseline is established, every future problem becomes easier to diagnose, because the user already knows what normal looks like for that specific machine in that specific home or office.
Open System Monitor right now while the machine is idle and write down three numbers: the percentage of CPU used, the percentage of memory used, and the percentage of the root partition that is full. Keep that note for a week, then come back and compare what the same three numbers look like a week later. The pattern they form is the first real diagnostic chart the machine has ever produced, and from there every slow day afterwards will have a clearer answer waiting in the same three tabs.
