l / streamlit_app /utils /performance.py
Princess3's picture
Upload 25 files
c089ca4 verified
raw
history blame
10.3 kB
#!/usr/bin/env python3
"""
Performance Monitor
Monitors system performance metrics for the NZ Legislation Loophole Analysis application.
Tracks memory usage, CPU utilization, processing times, and other performance indicators.
"""
import time
import threading
import psutil
from typing import Dict, Any, Optional, List
from collections import deque
import streamlit as st
class PerformanceMonitor:
"""Performance monitoring system"""
def __init__(self, max_history: int = 1000):
"""
Initialize performance monitor
Args:
max_history: Maximum number of historical data points to keep
"""
self.max_history = max_history
self.lock = threading.RLock()
# Historical data storage
self.memory_history = deque(maxlen=max_history)
self.cpu_history = deque(maxlen=max_history)
self.processing_times = deque(maxlen=max_history)
# Current metrics
self.current_metrics = {
'memory_usage_mb': 0,
'memory_percent': 0,
'cpu_percent': 0,
'active_threads': 0,
'processing_time_avg': 0,
'processing_time_max': 0,
'processing_time_min': 0,
'total_processed_chunks': 0,
'chunks_per_second': 0
}
# Processing timing
self.processing_start_time = None
self.last_chunk_time = time.time()
# Start monitoring thread
self.monitoring = True
self.monitor_thread = threading.Thread(target=self._monitor_loop, daemon=True)
self.monitor_thread.start()
def _monitor_loop(self):
"""Background monitoring loop"""
while self.monitoring:
try:
self._update_metrics()
time.sleep(1) # Update every second
except Exception as e:
print(f"Performance monitoring error: {e}")
time.sleep(5) # Wait longer on error
def _update_metrics(self):
"""Update current performance metrics"""
process = psutil.Process()
with self.lock:
# Memory metrics
memory_info = process.memory_info()
memory_usage_mb = memory_info.rss / 1024 / 1024
memory_percent = process.memory_percent()
# CPU metrics
cpu_percent = process.cpu_percent(interval=0.1)
# Thread count
active_threads = len(process.threads())
# Update current metrics
self.current_metrics.update({
'memory_usage_mb': memory_usage_mb,
'memory_percent': memory_percent,
'cpu_percent': cpu_percent,
'active_threads': active_threads
})
# Store historical data
current_time = time.time()
self.memory_history.append((current_time, memory_usage_mb))
self.cpu_history.append((current_time, cpu_percent))
def start_processing_timer(self):
"""Start timing a processing operation"""
self.processing_start_time = time.time()
def end_processing_timer(self) -> float:
"""End timing and return elapsed time"""
if self.processing_start_time is None:
return 0
elapsed = time.time() - self.processing_start_time
self.processing_start_time = None
with self.lock:
self.processing_times.append(elapsed)
# Update processing time statistics
if self.processing_times:
self.current_metrics['processing_time_avg'] = sum(self.processing_times) / len(self.processing_times)
self.current_metrics['processing_time_max'] = max(self.processing_times)
self.current_metrics['processing_time_min'] = min(self.processing_times)
return elapsed
def record_chunk_processing(self):
"""Record that a chunk has been processed"""
current_time = time.time()
with self.lock:
self.current_metrics['total_processed_chunks'] += 1
# Calculate chunks per second
time_diff = current_time - self.last_chunk_time
if time_diff > 0:
current_cps = 1.0 / time_diff
# Smooth the chunks per second calculation
self.current_metrics['chunks_per_second'] = (
0.9 * self.current_metrics['chunks_per_second'] + 0.1 * current_cps
)
self.last_chunk_time = current_time
def get_stats(self) -> Dict[str, Any]:
"""Get current performance statistics"""
with self.lock:
return self.current_metrics.copy()
def get_memory_history(self, time_window_seconds: int = 300) -> List[tuple]:
"""Get memory usage history within time window"""
current_time = time.time()
cutoff_time = current_time - time_window_seconds
with self.lock:
return [(t, v) for t, v in self.memory_history if t >= cutoff_time]
def get_cpu_history(self, time_window_seconds: int = 300) -> List[tuple]:
"""Get CPU usage history within time window"""
current_time = time.time()
cutoff_time = current_time - time_window_seconds
with self.lock:
return [(t, v) for t, v in self.cpu_history if t >= cutoff_time]
def get_processing_time_stats(self) -> Dict[str, Any]:
"""Get processing time statistics"""
with self.lock:
if not self.processing_times:
return {
'count': 0,
'average': 0,
'maximum': 0,
'minimum': 0,
'median': 0
}
sorted_times = sorted(self.processing_times)
return {
'count': len(self.processing_times),
'average': sum(self.processing_times) / len(self.processing_times),
'maximum': max(self.processing_times),
'minimum': min(self.processing_times),
'median': sorted_times[len(sorted_times) // 2]
}
def get_system_info(self) -> Dict[str, Any]:
"""Get system information"""
return {
'cpu_count': psutil.cpu_count(),
'cpu_count_logical': psutil.cpu_count(logical=True),
'total_memory_gb': psutil.virtual_memory().total / (1024**3),
'available_memory_gb': psutil.virtual_memory().available / (1024**3),
'python_version': f"{psutil.python_implementation()} {psutil.python_version()}",
'platform': psutil.platform
}
def reset_stats(self):
"""Reset performance statistics"""
with self.lock:
self.processing_times.clear()
self.current_metrics['total_processed_chunks'] = 0
self.current_metrics['chunks_per_second'] = 0
self.current_metrics['processing_time_avg'] = 0
self.current_metrics['processing_time_max'] = 0
self.current_metrics['processing_time_min'] = 0
def cleanup(self):
"""Cleanup resources"""
self.monitoring = False
if self.monitor_thread.is_alive():
self.monitor_thread.join(timeout=2)
def get_performance_report(self) -> Dict[str, Any]:
"""Generate a comprehensive performance report"""
return {
'current_metrics': self.get_stats(),
'processing_stats': self.get_processing_time_stats(),
'system_info': self.get_system_info(),
'memory_history_count': len(self.memory_history),
'cpu_history_count': len(self.cpu_history),
'processing_times_count': len(self.processing_times)
}
def check_memory_threshold(self, threshold_mb: int) -> bool:
"""Check if memory usage is above threshold"""
return self.current_metrics['memory_usage_mb'] > threshold_mb
def check_cpu_threshold(self, threshold_percent: float) -> bool:
"""Check if CPU usage is above threshold"""
return self.current_metrics['cpu_percent'] > threshold_percent
def get_recommendations(self) -> List[str]:
"""Get performance recommendations based on current metrics"""
recommendations = []
# Memory recommendations
if self.current_metrics['memory_usage_mb'] > 7000:
recommendations.append("High memory usage detected. Consider reducing batch size or chunk size.")
elif self.current_metrics['memory_usage_mb'] > 5000:
recommendations.append("Moderate memory usage. Monitor closely during processing.")
# CPU recommendations
if self.current_metrics['cpu_percent'] > 90:
recommendations.append("High CPU usage. Consider reducing processing intensity.")
elif self.current_metrics['cpu_percent'] > 70:
recommendations.append("Moderate CPU usage. Processing is running optimally.")
# Processing speed recommendations
avg_time = self.current_metrics.get('processing_time_avg', 0)
if avg_time > 10:
recommendations.append("Slow processing detected. Consider using a more powerful model or optimizing settings.")
elif avg_time > 5:
recommendations.append("Moderate processing speed. Consider increasing batch size if memory allows.")
# Cache recommendations
# This would be integrated with cache manager stats
chunks_per_second = self.current_metrics.get('chunks_per_second', 0)
if chunks_per_second < 1:
recommendations.append("Low processing throughput. Consider optimizing chunk size or model parameters.")
if not recommendations:
recommendations.append("Performance is optimal. All metrics are within normal ranges.")
return recommendations
# Global performance monitor instance
_performance_instance = None
_performance_lock = threading.Lock()
def get_performance_monitor(max_history: int = 1000) -> PerformanceMonitor:
"""Get or create global performance monitor instance"""
global _performance_instance
with _performance_lock:
if _performance_instance is None:
_performance_instance = PerformanceMonitor(max_history)
return _performance_instance