docs: update reference.md

Signed-off-by: Rodney Osodo <socials@rodneyosodo.com>

Author: rodneyosodo

docs/monitoring.md

@@ -4,7 +4,7 @@
 
 Propeller tracks OS-level metrics for every WebAssembly task running on proplets. The system collects CPU usage, memory consumption, disk I/O, thread counts, and more with minimal performance overhead.
 
-Both Go and Rust proplets include built-in monitoring using `gopsutil` for Go and `sysinfo` for Rust. These libraries provide cross-platform compatibility across Linux, macOS, and Windows environments.
+Propeller tracks OS-level metrics for every WebAssembly task running on proplets. The system collects CPU usage, memory consumption, disk I/O, thread counts, and more with minimal performance overhead. The Rust proplet includes built-in monitoring using the `sysinfo` crate for cross-platform metrics. This provides compatibility across Linux, macOS, and Windows environments.
 
 ## Architecture
 
@@ -16,7 +16,9 @@ Each proplet monitors its running tasks independently and reports metrics to the
 
 1. Proplet spawns a monitoring thread for each task
 2. Thread collects process metrics at configured intervals (1-120 seconds)
-3. Metrics are published to MQTT topic `m/{domain}/c/{channel}/metrics/proplet`
+3. Metrics are published to MQTT topics:
+   - Proplet-level: `m/{domain}/c/{channel}/control/proplet/metrics`
+   - Task-level: `m/{domain}/c/{channel}/control/proplet/task_metrics`
 4. Manager receives and stores metrics in memory
 5. API clients query metrics via HTTP endpoints
 

docs/reference.md

@@ -1,6 +1,6 @@
 # Process Monitoring Implementation
 
-This document describes the complete process monitoring implementation for both Go and Rust proplets in the Propeller distributed task execution system.
+This document describes the complete process monitoring implementation for Rust proplets in the Propeller distributed task execution system.
 
 ## Overview
 
@@ -12,18 +12,29 @@ Comprehensive OS-level process monitoring has been implemented for:
 
 ## Monitoring Profiles
 
-Both implementations provide identical profiles:
+Profiles define which metrics to collect, how often, and how much history to retain.
+
+
+The Rust implementation provides two built-in profiles:
 
 | Profile             | Interval | Metrics                            | Export | History | Use Case            |
 | ------------------- | -------- | ---------------------------------- | ------ | ------- | ------------------- |
 | Standard            | 10s      | All                                | Yes    | 100     | General purpose     |
-| Minimal             | 60s      | CPU, Memory                        | No     | 0       | Lightweight         |
-| Intensive           | 1s       | All                                | Yes    | 1000    | Debug/analysis      |
-| Batch Processing    | 30s      | CPU, Memory, Disk                  | Yes    | 200     | Data processing     |
-| Real-time API       | 5s       | CPU, Memory, Network, Threads, FDs | Yes    | 500     | HTTP/API servers    |
 | Long-running Daemon | 120s     | All                                | Yes    | 500     | Background services |
-| Disabled            | -        | None                               | No     | 0       | No monitoring       |
-
+### Custom Profiles
+
+You can also define custom monitoring profiles via JSON configuration with the following options:
+
+- `enabled`: Enable/disable monitoring (default: `true`)
+- `interval`: Collection interval in seconds (default: `10`)
+- `collect_cpu`: Collect CPU metrics (default: `true`)
+- `collect_memory`: Collect memory metrics (default: `true`)
+- `collect_disk_io`: Collect disk I/O metrics (default: `true`)
+- `collect_threads`: Collect thread count (default: `true`)
+- `collect_file_descriptors`: Collect file descriptor count (default: `true`)
+- `export_to_mqtt`: Publish metrics to MQTT (default: `true`)
+- `retain_history`: Keep metrics history (default: `true`)
+- `history_size`: Maximum history entries (default: `100`)
 ## Metrics Collected
 
 ### Common Metrics (All Platforms)
@@ -44,11 +55,10 @@ Both implementations provide identical profiles:
 
 ## MQTT Topics
 
-### Proplet-Level Metrics (Go)
+### Proplet-Level Metrics
 
 ```txt
 m/{domain_id}/c/{channel_id}/control/proplet/metrics
-```
 
 Publishes overall proplet health metrics.
 
@@ -96,18 +106,12 @@ Publishes per-task process metrics.
 
 ## Configuration
 
-### Go Proplet Environment Variables
-
-```bash
-PROPLET_METRICS_INTERVAL=10        # Interval in seconds
-PROPLET_ENABLE_MONITORING=true     # Enable/disable
-```
 
 ### Rust Proplet Environment Variables
 
 ```bash
-PROPLET_ENABLE_MONITORING=true     # Enable/disable
-PROPLET_METRICS_INTERVAL=10        # Interval in seconds
+export PROPLET_ENABLE_MONITORING=true     # Enable/disable monitoring (default: true)
+export PROPLET_METRICS_INTERVAL=10        # Proplet-level metrics interval in seconds (default: 10)
 ```
 
 ### Per-Task Configuration (JSON)
@@ -142,7 +146,6 @@ Measured overhead across platforms:
 
 ## Usage Examples
 
-### Go - Start Task with Monitoring
 
 ```go
 task := task.Task{