Knowing what your VPS is actually doing takes about thirty seconds, and it is the difference between guessing that you need a bigger plan and knowing it. Windows ships with everything you need.
Key Takeaways
- Task Manager answers “is something wrong right now” in a few seconds.
- Resource Monitor answers “which process is responsible”.
- Memory usually runs out before CPU on a trading VPS. Budget 300-500 MB per MT4 terminal.
- Judge sustained load, not spikes. Brief peaks during news are normal.
1. Task Manager: The Thirty-Second Check
Connect over RDP and press Ctrl+Shift+Esc. Open the Performance tab.
You get live graphs for CPU, Memory, Disk and Ethernet. This is where you start, because it tells you immediately whether the machine is under pressure and which of the four is the constraint.
Read it while your terminals are running the way they normally do. A reading taken right after boot, before the platforms have loaded history and connected, tells you very little.
2. Resource Monitor: Which Process Is Responsible
Task Manager tells you that something is consuming resources. Resource Monitor tells you what.
From the Performance tab, click Open Resource Monitor at the bottom. The CPU, Memory and Disk tabs each list processes sorted by usage, so you can see which specific terminal is the heavy one rather than guessing.
This matters when you run several terminals. It is common for one instance carrying a dozen charts and a heavy indicator set to account for most of the load while the others sit nearly idle.
3. What the Numbers Should Look Like
Rough guidance for a working trading VPS:
- CPU — idle terminals sit low. Brief spikes to 100% when a terminal loads history or recalculates an indicator are normal. Sustained load above 80% during ordinary market hours is not.
- Memory — roughly 300-500 MB per MT4 terminal, more with many charts or a long history. This is usually the first thing to run out.
- Disk — should be near idle most of the time. Sustained 100% activity is almost always Windows Update, an antivirus scan, or a terminal rebuilding history.
Judge all three on sustained behaviour during normal hours, not on a single reading during a news release.
4. What to Do About It
If you are consistently over those thresholds, work through this order:
- Close charts you are not using. Every open chart holds bars in memory and recalculates indicators. This is the single biggest saving available.
- Hide unused symbols in Market Watch. Every visible symbol receives ticks and costs bandwidth and CPU.
- Reduce max bars in history and chart, to comfortably more than your longest EA lookback rather than to the minimum.
- Close terminals you do not need running. Two terminals at the same broker means two full copies in memory.
- Then upgrade, once you have done the above and the machine is still under pressure.
5. Watching It Over Time
Task Manager only shows you the present moment. For a longer view, open Performance Monitor (perfmon from the Start menu or Run box) and add counters for Processor > % Processor Time, Memory > Available MBytes and PhysicalDisk > % Disk Time.
This is worth doing when the problem is intermittent, for example a slowdown that only shows up at the New York open or during a specific EA’s trading window, and you are never watching at the moment it happens.
Need Help?
Support is included free with every FXVPS plan. If your VPS feels slow and the numbers above do not explain why, reach out via live chat on fxvps.biz or email support@fxvps.biz with what you are seeing and how many terminals you are running.