Visualizing Zabbix Metrics with Grafana (Part 3) — Building a Static Dashboard for CPU, Memory, and Disk

📘 Overview

After completing the Zabbix integration, it’s time to move on to the dashboard design phase.
This guide goes beyond simply importing templates — it walks through how to create a practical Grafana dashboard that allows operators to interpret and correlate key metrics such as CPU, memory, disk, and network performance.

A dashboard isn’t just a collection of graphs — it’s an operational map that visually guides you on where to start investigating when a problem occurs.

1. Create Folders and Dashboards

1.1 Creating a Folder (Optional)

When managing multiple server groups or metric sets, using folders helps keep your dashboards organized.

Menu → Dashboards → New → New folder
→ Example name: example → Create

1.2 Creating a Dashboard

Inside the folder, create a new dashboard.

Menu → Dashboards → example → New dashboard → Add visualization

2. First Panel: CPU Utilization

2.1 Basic Query Settings

2.2 Visualization Options

2.3 Why These Metrics Matter

⚙️ Recommended additional metrics for advanced or virtualized environments

• CPU steal time → Detects hypervisor resource contention
• CPU nice time → Indicates low-priority process share
• CPU softirq time → Identifies network interrupt concentration

3. Choosing the Correct Value Function (Last vs. Last*)

Grafana’s Zabbix data often includes Null values.
To display data correctly, you must choose the proper “last value function.”

✅ In production, using Last* is recommended.
It preserves the last valid value even if data gaps occur, preventing visual interruptions.

4. Building Logical Relationships Between Metrics

A good dashboard doesn’t just list metrics — it shows the logical flow of investigation:
“Symptom → Cause analysis → Root cause.”

Example flow 👇

CPU panel → Detect performance issue
    ↓
Disk I/O panel → Check for bottlenecks
    ↓
Memory/Process panel → Identify resource culprits
    ↓
Network panel → Examine external influence

By structuring panels this way, operators can see the flow of system anomalies at a glance.
In other words, meaningful relationships matter more than the number of graphs.

5. Saving the Dashboard

Once configuration is complete:
Save dashboard → Enter title → Save

📌 Example title: example server
You can later expand this same structure for Memory, Disk, and Network dashboards.

6. Verification & Validation

7. Next Step

In the next part, we’ll use Grafana Variables to build dynamic dashboards — allowing flexible selection of Group, Host, and Item for interactive metric views.

🛠 마지막 수정일: 2025.11.10