Documentation
Quick start + feature guide for every section and device (continuously updated)
Health & Dashboard
This article introduces the dashboard's server health monitoring features and the meaning of each metric.
Overview
The dashboard displays the real-time health status of all servers at once: CPU load, memory usage, disk usage, system information, and online status. Open the panel home page to see all servers at a glance.
Use Cases
- Quickly tell which servers are running normally and which have anomalies
- Monitor whether memory/disk is nearly full, for early warning
- View a server's basic information (CPU model, core count, OS version, architecture)
What Is Displayed (What Each Metric Means)
Each server card shows:
| Metric | Meaning | Plain explanation |
|---|---|---|
| Online status | Online / Offline (green dot / red dot) | Whether the panel can reach it over SSH |
| CPU load | System load average | The closer the number is to the "core count", the busier it is; sustained load above core count means the CPU is strained |
| Memory | Total / used / available | Too little available memory (<8%) warns of OOM risk |
| Disk | Root partition total / used / usage | Usage above 90% is time to clean up |
| Uptime | Days since last reboot | A large value means the system is stable |
| CPU model / cores / architecture | e.g. "4 cores / x86_64 / Intel Xeon" | Hardware configuration |
| Operating system | Distribution name and version | e.g. "Ubuntu 22.04" |
| Hostname | The server's hostname | Machine name |
| LAN IP | The server's internal address | Internal address |
The Relationship Between Load and Core Count
- Load < core count: normal, the machine has idle capacity
- Load = core count: exactly saturated
- Load > core count: overloaded, may become slow
- Load > core count × 3: severely overloaded (Crash Prevention intervenes at this threshold by default)
Refresh Mechanism
- The panel checks all servers in batch (in parallel), avoiding the slowness of connecting one by one
- Health data has a 60-second cache: refreshes within the same minute will not repeat SSH connections
- Servers that fail detection have a 30-second failure cache: offline servers won't waste a timeout wait on every refresh
> Therefore, when you manually click "Refresh" and the data doesn't change, it is normal caching behavior — it will update automatically after 60 seconds.
Online / Offline Status
"Offline" does not mean the server is down — it only means "the panel cannot reach this server over SSH". Common causes:
1. SSH port unreachable (cloud firewall / security group blocking, expired NAT port mapping)
2. Wrong username or password
3. The server is actually down
See the "Server shows Offline" section of the FAQ for troubleshooting.
Notes
1. Health checks are performed by running commands remotely over SSH — nothing is installed on the target server and no target resources are consumed.
2. For "tunnel servers" (no public IP), health data is transferred over the reverse tunnel; the detection method is the same as for ordinary servers.
3. If a server has been offline for a long time and is no longer in use, consider deleting it in "Server Management" to avoid waiting for a timeout on every refresh.