ElBruno.OllamaMonitor

Configuration Guide

Overview

ElBruno.OllamaMonitor stores configuration in a JSON file that you can edit directly or modify via CLI commands.

Configuration File Location

%LOCALAPPDATA%\ElBruno\OllamaMonitor\settings.json

On Windows:

• %LOCALAPPDATA% typically expands to C:\Users\<YourUsername>\AppData\Local
• So the full path is usually: C:\Users\<YourUsername>\AppData\Local\ElBruno\OllamaMonitor\settings.json

The file is created automatically with default values on first run.

Default Configuration

{
  "endpoint": "http://localhost:11434",
  "refreshIntervalSeconds": 2,
  "startMinimizedToTray": true,
  "showFloatingWindowOnStart": false,
  "enableGpuMetrics": true,
  "enableDiskMetrics": true,
  "highCpuThresholdPercent": 80,
  "highMemoryThresholdGb": 16,
  "highGpuThresholdPercent": 85
}

Configuration Options

endpoint

Type: string
Default: http://localhost:11434

The HTTP endpoint where Ollama is running. Modify this if:

• Ollama is running on a different port
• Ollama is running on a different machine (e.g., VM, remote server)

Examples:

"endpoint": "http://localhost:11434"      // Local default
"endpoint": "http://192.168.1.100:11434"  // Remote machine
"endpoint": "http://ollama.local:11434"   // DNS name

refreshIntervalSeconds

Type: int (1 or greater)
Default: 2

How often (in seconds) the app polls the Ollama API and system metrics.

• Lower values (1–2s): More responsive, slightly more CPU
• Higher values (5–10s): Less responsive, lighter load

Examples:

"refreshIntervalSeconds": 1   // Frequent polling
"refreshIntervalSeconds": 5   // Moderate polling
"refreshIntervalSeconds": 10  // Low frequency

startMinimizedToTray

Type: bool
Default: true

Whether the app starts minimized to the system tray (hidden from desktop).

• true: App launches hidden; use the tray icon to open the details window or mini monitor
• false: App launches with the details window visible

showFloatingWindowOnStart

Type: bool
Default: false

Whether to show the standard details window automatically when the app starts.

• Works independently of startMinimizedToTray
• Useful if you want the details visible on startup

enableGpuMetrics

Type: bool
Default: true

Whether to attempt to collect NVIDIA GPU metrics.

• Requires nvidia-smi to be available on PATH
• If not available, this setting has no effect (GPU data will show as “N/A”)
• Set to false to skip GPU polling entirely

enableDiskMetrics

Type: bool
Default: true

Whether to collect disk read/write metrics for the Ollama process.

• Currently best-effort; may show “N/A” on some Windows versions
• Set to false to disable disk metric collection

highCpuThresholdPercent

Type: double (0–100)
Default: 80

The CPU usage percentage threshold that triggers the Orange tray icon state (HighUsage).

When Ollama’s CPU usage exceeds this threshold and a model is loaded, the tray icon turns orange to indicate active, resource-intensive work.

Examples:

"highCpuThresholdPercent": 50   // Aggressive threshold
"highCpuThresholdPercent": 80   // Moderate (default)
"highCpuThresholdPercent": 95   // Conservative

highMemoryThresholdGb

Type: double
Default: 16

The RAM usage (in GB) threshold that triggers the Orange tray icon state.

When Ollama’s memory usage exceeds this threshold, the tray icon turns orange.

Examples:

"highMemoryThresholdGb": 4      // Tight constraint
"highMemoryThresholdGb": 16     // Moderate (default)
"highMemoryThresholdGb": 32     // Permissive

highGpuThresholdPercent

Type: double (0–100)
Default: 85

The GPU usage percentage threshold that triggers the Orange state.

When GPU utilization exceeds this threshold, the tray icon turns orange.

Examples:

"highGpuThresholdPercent": 70   // Aggressive
"highGpuThresholdPercent": 85   // Moderate (default)
"highGpuThresholdPercent": 99   // Conservative

Editing Configuration

Option 1: CLI Commands

In Phase 1, you can use ollamamon config commands to manage the endpoint and refresh interval:

# View current configuration
ollamamon config

# Change the Ollama endpoint
ollamamon config set endpoint http://192.168.1.100:11434

# Change refresh interval to 5 seconds
ollamamon config set refresh-interval 5

# Reset to default settings
ollamamon config reset

Note: To change thresholds, GPU metrics, or disk metrics settings, use Option 2 (direct file edit) below.

Option 2: Direct File Edit

1. Open the settings file in a text editor (Notepad, Visual Studio Code, etc.):

   %LOCALAPPDATA%\ElBruno\OllamaMonitor\settings.json
   

2. Edit the JSON values

3. Save the file

4. Restart the application for changes to take effect

Example:

{
  "endpoint": "http://192.168.1.50:11434",
  "refreshIntervalSeconds": 3,
  "startMinimizedToTray": true,
  "showFloatingWindowOnStart": false,
  "enableGpuMetrics": true,
  "enableDiskMetrics": true,
  "highCpuThresholdPercent": 75,
  "highMemoryThresholdGb": 12,
  "highGpuThresholdPercent": 80
}

Tray Icon States

The tray icon color reflects the current state, which is partly determined by your threshold settings:

State	Color	Trigger	CPU/RAM/GPU
NotReachable	Gray	Ollama API unreachable	—
Running	Green	API reachable, no model	Low
ModelLoaded	Blue	Model loaded	Low
HighUsage	Orange	Model running	Exceeds threshold
Error	Red	Unexpected error	—

The thresholds you set control when the app transitions from ModelLoaded to HighUsage.

Troubleshooting Configuration

Settings file is corrupted or missing

If the settings file is corrupted, the app will log an error and attempt to use defaults. To reset:

ollamamon config reset

This recreates the file with default values.

Changes don’t take effect immediately

Configuration changes require an app restart. Either:

• Close the app from the tray menu
• Run ollamamon again to launch a fresh instance

“Endpoint unreachable” message

If you see “Endpoint unreachable” in the floating window:

1. Verify Ollama is actually running: ollama serve
2. Check the endpoint setting: ollamamon config
3. Test connectivity: curl http://localhost:11434/api/version (or your configured endpoint)
4. If Ollama is on a different machine, use that IP/hostname instead

GPU metrics show “N/A”

This is normal if:

• NVIDIA GPU is not installed
• nvidia-smi is not installed or not on PATH
• GPU metrics are disabled in settings

To enable GPU metrics:

ollamamon config set gpu-metrics true

Then verify nvidia-smi is available:

nvidia-smi

If nvidia-smi is not found, install NVIDIA drivers from nvidia.com.

Advanced: Custom Refresh Interval

If you want to balance responsiveness and CPU usage:

• 1–2 seconds: Highly responsive but slightly more CPU (good for demo/presentation)
• 3–5 seconds: Reasonable balance
• 10+ seconds: Light-weight monitoring (good for background use)

ollamamon config set refresh-interval 10

Advanced: Remote Ollama Monitoring

To monitor Ollama running on another machine:

1. Ensure Ollama is accessible from your machine (firewall rules, etc.)
2. Set the endpoint to the remote machine:

   ollamamon config set endpoint http://<remote-ip>:11434
   

3. Launch the app: ollamamon

Note: Remote monitoring is best-effort in Phase 1. For production use, consider Phase 2 features.

---

Next Steps:

• Development Guide — Build and modify the app
• Troubleshooting — Common issues and fixes
• Architecture Guide — Technical deep dive

This site is open source. Improve this page.