Files
cbc2027/services/metrics_collector.py

138 lines
4.6 KiB
Python

"""
Performance metrics collector for CBCFacil
"""
import time
import threading
import psutil
import logging
from typing import Dict, Any, Optional
from datetime import datetime, timedelta
from contextlib import contextmanager
class MetricsCollector:
"""Collect and aggregate performance metrics"""
def __init__(self):
self.logger = logging.getLogger(__name__)
self._start_time = time.time()
self._request_count = 0
self._error_count = 0
self._total_latency = 0.0
self._latencies = []
self._lock = threading.Lock()
self._process = psutil.Process()
def record_request(self, latency: float, success: bool = True) -> None:
"""Record a request with latency"""
with self._lock:
self._request_count += 1
self._total_latency += latency
self._latencies.append(latency)
# Keep only last 1000 latencies for memory efficiency
if len(self._latencies) > 1000:
self._latencies = self._latencies[-1000:]
if not success:
self._error_count += 1
def get_latency_percentiles(self) -> Dict[str, float]:
"""Calculate latency percentiles"""
with self._lock:
if not self._latencies:
return {"p50": 0, "p95": 0, "p99": 0}
sorted_latencies = sorted(self._latencies)
n = len(sorted_latencies)
return {
"p50": sorted_latencies[int(n * 0.50)],
"p95": sorted_latencies[int(n * 0.95)],
"p99": sorted_latencies[int(n * 0.99)]
}
def get_system_metrics(self) -> Dict[str, Any]:
"""Get system resource metrics"""
try:
memory = self._process.memory_info()
cpu_percent = self._process.cpu_percent(interval=0.1)
return {
"cpu_percent": cpu_percent,
"memory_rss_mb": memory.rss / 1024 / 1024,
"memory_vms_mb": memory.vms / 1024 / 1024,
"thread_count": self._process.num_threads(),
"open_files": self._process.open_files(),
}
except Exception as e:
self.logger.warning(f"Error getting system metrics: {e}")
return {}
def get_summary(self) -> Dict[str, Any]:
"""Get metrics summary"""
with self._lock:
uptime = time.time() - self._start_time
latency_pcts = self.get_latency_percentiles()
return {
"uptime_seconds": round(uptime, 2),
"total_requests": self._request_count,
"error_count": self._error_count,
"error_rate": round(self._error_count / max(1, self._request_count) * 100, 2),
"requests_per_second": round(self._request_count / max(1, uptime), 2),
"average_latency_ms": round(self._total_latency / max(1, self._request_count) * 1000, 2),
"latency_p50_ms": round(latency_pcts["p50"] * 1000, 2),
"latency_p95_ms": round(latency_pcts["p95"] * 1000, 2),
"latency_p99_ms": round(latency_pcts["p99"] * 1000, 2),
}
def reset(self) -> None:
"""Reset metrics"""
with self._lock:
self._request_count = 0
self._error_count = 0
self._total_latency = 0.0
self._latencies = []
self._start_time = time.time()
class LatencyTracker:
"""Context manager for tracking operation latency"""
def __init__(self, collector: MetricsCollector, operation: str):
self.collector = collector
self.operation = operation
self.start_time: Optional[float] = None
self.success = True
def __enter__(self):
self.start_time = time.time()
return self
def __exit__(self, exc_type, exc_val, exc_tb):
latency = time.time() - self.start_time
success = exc_type is None
self.collector.record_request(latency, success)
return False # Don't suppress exceptions
# Global metrics collector
metrics_collector = MetricsCollector()
@contextmanager
def track_latency(operation: str = "unknown"):
"""Convenience function for latency tracking"""
with LatencyTracker(metrics_collector, operation):
yield
def get_performance_report() -> Dict[str, Any]:
"""Generate comprehensive performance report"""
return {
"metrics": metrics_collector.get_summary(),
"system": metrics_collector.get_system_metrics(),
"timestamp": datetime.utcnow().isoformat()
}