Skip to content

README-MONITORING-DEBUGGING.md

Latest commit

 

History

History
113 lines (84 loc) · 2.95 KB

README-MONITORING-DEBUGGING.md

File metadata and controls

113 lines (84 loc) · 2.95 KB
Raw

from pathlib import Path

readme_content = """

🛠️ Monitoring and Debugging - Ms. Potts MLOps

This document outlines the setup and usage of the monitoring and debugging modules used in the Ms. Potts AI assistant project. These modules are designed to provide system health insights and help trace and resolve errors efficiently during development and deployment.

📡 Monitoring

✅ Features:

  • Periodic system metrics logging (CPU, Memory, Disk)
  • Application-level metrics logging (latency, status codes, endpoint access)

📦 Module:

src/ms_potts/utils/monitoring.py

📁 Output Directory:

metrics/

🧪 How It Works:

Monitoring is enabled via the ModelMonitor class.

from utils.monitoring import ModelMonitor
monitor = ModelMonitor(metrics_dir="./metrics")
monitor.start_monitoring(interval=5)  # Logs system stats every 5s

📊 Example Metric Log:

{
  "timestamp": "2025-05-22T12:00:00",
  "cpu_usage": 25.4,
  "memory_usage": 63.1,
  "disk_usage": 2.9
}

System Monitoring Visualization

🧠 Application Metrics Example:

monitor.log_application_metrics({
    "endpoint": "/query",
    "status_code": 200,
    "processing_time_ms": 123.4
})

🐞 Debugging

✅ Features:

  • Trace every key step and intermediate value
  • Logs function calls with input/output
  • Saves execution traces and error traces automatically

📦 Module:

src/ms_potts/utils/debugging.py

📁 Output Directory:

debug_traces/

⚙️ How It Works:

Use the DebugTracer to trace functions and capture debug values.

from utils.debugging import DebugTracer, debug_value

tracer = DebugTracer(output_dir="./debug_traces")

@tracer.trace_function
def get_response(...):
    debug_value(query_embedding, "Query Embedding Shape")

📄 Output Files:

  • query_trace_<timestamp>.json: for successful traces
  • error_trace_<timestamp>.json: for failed runs

📘 Example Trace Log:

{
  "function": "get_response",
  "step": "intent_classification",
  "value": "Meal-Logging"
}

📌 Deliverables Checklist

Component Path Status
Monitoring module src/ms_potts/utils/monitoring.py ✅
Debugging module src/ms_potts/utils/debugging.py ✅
Metrics samples metrics/ ✅
Trace logs debug_traces/ ✅
This documentation README_MONITORING_DEBUGGING.md ✅

📎 Notes

  • Logging integrates with EnhancedLogger for rich console + file output.
  • Future enhancements can include Prometheus/Grafana integration for live dashboards.

🧪 Tested & Verified: Locally and within Docker container during API queries. """