So your PRTG CPU graph refuses to sit nicely at 100% scale. Instead, it auto-adjusts. It stretches. It shrinks. And suddenly your clean monitoring dashboard looks… weird. Don’t worry. Fixing this is easier than you think.
TLDR: By default, PRTG auto-scales CPU graphs based on usage. To lock the graph at 100%, you need to modify the channel settings and turn off automatic scaling. Set a custom maximum value of 100 for the CPU channel. Once done, your graph will always show usage against a stable 0–100% scale, making spikes and trends easier to spot.
Now let’s break this down step by step. Simple. Clear. No stress.
Why Set the CPU Graph to 100%?
PRTG loves automation. It automatically adjusts graph scaling to match your data. That sounds good. Until it isn’t.
Here’s the problem:
- If CPU usage peaks at 40%, the graph may scale to 50%.
- If it peaks at 12%, the graph may scale to 15%.
- Small variations look huge.
That can be misleading.
You might think your server is struggling. But really, it’s just a stretched graph.
By forcing the CPU graph to always display from 0% to 100%, you:
- Get consistent scaling.
- Compare servers more easily.
- Spot real spikes instantly.
- Avoid unnecessary panic.
Consistency is key in monitoring. Especially when you manage multiple devices.
How PRTG Graph Scaling Works
Before adjusting settings, let’s quickly understand what’s happening.
PRTG channels use something called:
- Automatic scaling
- Manual scaling
By default, most CPU sensors use automatic scaling. This means:
- PRTG checks the highest data point.
- It sets the Y-axis slightly above that value.
- The graph visually fills the space.
That makes graphs look dynamic. But not standardized.
And standardized is exactly what we want.
Step-by-Step: Set PRTG CPU Graph to 100% Scale
Let’s walk through the process.
These steps work for:
- Windows CPU sensors
- SNMP CPU Load sensors
- WMI CPU sensors
Ready? Let’s go.
Step 1: Open Your CPU Sensor
Log into your PRTG web interface.
Navigate to:
- Your device
- Select the CPU sensor
You’ll now see the sensor overview page.
Image not found in postmetaThis page shows live usage, historic graphs, and channel details.
Step 2: Go to Channel Settings
CPU sensors often have one or more channels. Most commonly:
- Total CPU
- CPU Core 1, 2, 3, etc.
Click the channel name. For example, click Total.
This opens the channel settings page.
Scroll down until you see Display Settings.
Step 3: Disable Automatic Scaling
This is the key part.
Look for the setting called:
Vertical Axis Scaling
By default, it is set to:
- Automatic
Change it to:
- Manual
Once switched, additional fields will appear.
You will now see fields for:
- Minimum
- Maximum
Step 4: Set the Maximum to 100
This is where the magic happens.
Set:
- Minimum = 0
- Maximum = 100
Simple.
Click Save.
Your CPU graph is now locked to a 0–100% scale.
What If You Have Multiple CPU Channels?
Good question.
If your sensor tracks multiple cores, you may have:
- CPU Core 1
- CPU Core 2
- CPU Core 3
- And so on
Each channel can have its own scaling settings.
If you want all graphs standardized:
- Repeat the same process for each channel.
Yes. It takes a few minutes.
But it keeps your graphs clean and consistent.
Before and After: What Changes?
Here’s what happens visually.
Before (Automatic Scaling)
- Graph stretches to highest recorded value.
- Small CPU jumps look dramatic.
- Comparisons between devices are hard.
After (Manual 100% Scaling)
- Graph always shows true percentage scale.
- Spikes are properly proportioned.
- Multiple servers look consistent.
The difference is subtle. But powerful.
Extra Tip: Adjusting Historic Report Graphs
Sometimes, users say:
“My live graph is fixed. But my reports still auto-scale.”
This can happen depending on report templates.
In that case:
- Go to Reports
- Edit your report settings
- Check graph scaling options
Some reports allow independent scaling configurations.
Make sure they align with your manual channel settings.
When Should You NOT Use 100% Scaling?
Let’s be fair.
Manual scaling isn’t always perfect.
You might prefer automatic scaling if:
- You monitor very low-load systems.
- You care about tiny usage changes under 5%.
- You’re troubleshooting micro-spikes.
Automatic scaling can exaggerate very small changes. And that can sometimes be helpful.
It depends on your goals.
Common Mistakes to Avoid
Let’s save you some frustration.
1. Forgetting to Save
It happens.
You change settings. Navigate away. Nothing updates.
Always click Save.
2. Editing the Wrong Channel
Make sure you're adjusting the correct channel.
Sometimes “Total CPU” and “CPU Load” are separate sensors.
3. Not Checking Inheritance Settings
PRTG uses inheritance.
If settings are inherited from parent objects, you may need to:
- Disable inheritance
- Enable custom settings
Look for the small lock icons next to settings.
Quick Comparison: Automatic vs Manual Scaling
| Feature | Automatic Scaling | Manual 0–100% Scaling |
|---|---|---|
| Visual Clarity | Dynamic but inconsistent | Stable and standardized |
| Easy Server Comparison | No | Yes |
| Spotting Minor Changes | Excellent | Moderate |
| Best For | Micro analysis | Performance overview |
| Risk of Misinterpretation | High | Low |
If you manage multiple servers, manual scaling usually wins.
Pro Move: Standardize All Performance Sensors
Why stop at CPU?
You can apply the same logic to:
- Memory usage
- Disk usage
- Bandwidth utilization
Setting consistent maximum values makes your dashboard:
- Cleaner
- More professional
- Easier to explain to clients
Especially during meetings.
No more “Why does this server look worse?” questions.
Final Thoughts
Setting your PRTG CPU graph to 100% scale is a small change.
But it makes a big difference.
You gain:
- Clarity
- Consistency
- Confidence in what you’re seeing
Monitoring should reduce stress. Not create it.
Locking your CPU graph to 0–100% ensures you’re measuring reality. Not graphical exaggeration.
Now your dashboard tells the truth.
And that’s exactly what good monitoring is all about.
