sheikh-kitty / monitoring /monitoring.py
likhonsheikh's picture
Upload folder using huggingface_hub
0efaf6e verified
"""
Sheikh-Kitty Monitoring System
Real-time metrics aggregation and system health monitoring
Features:
- API request metrics tracking
- Sandbox execution monitoring
- System resource monitoring
- Security violation alerts
- Performance analytics
- Health check endpoints
Author: MiniMax Agent
Date: 2025-11-14
"""
import json
import time
import psutil
import threading
import queue
from datetime import datetime, timedelta
from pathlib import Path
from typing import Dict, List, Optional, Any, Callable
from dataclasses import dataclass, asdict
from enum import Enum
import logging
import statistics
from collections import deque, defaultdict
import os
# Configure logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
class MetricType(Enum):
"""Types of metrics to track"""
COUNTER = "counter"
GAUGE = "gauge"
HISTOGRAM = "histogram"
TIMER = "timer"
class AlertSeverity(Enum):
"""Alert severity levels"""
INFO = "info"
WARNING = "warning"
ERROR = "error"
CRITICAL = "critical"
@dataclass
class Metric:
"""Individual metric data point"""
name: str
value: float
metric_type: MetricType
timestamp: datetime
labels: Dict[str, str] = None
tags: List[str] = None
@dataclass
class Alert:
"""System alert"""
id: str
severity: AlertSeverity
message: str
timestamp: datetime
metric_name: str
threshold: float
current_value: float
resolved: bool = False
resolved_at: Optional[datetime] = None
class MetricCollector:
"""Collect and store metrics"""
def __init__(self, max_history: int = 10000):
self.max_history = max_history
self.metrics = deque(maxlen=max_history)
self.current_values = {} # For gauge metrics
self.counters = defaultdict(float) # For counter metrics
self.lock = threading.Lock()
def record(self, metric: Metric):
"""Record a metric"""
with self.lock:
self.metrics.append(metric)
# Update current values for gauge metrics
if metric.metric_type == MetricType.GAUGE:
self.current_values[metric.name] = metric.value
elif metric.metric_type == MetricType.COUNTER:
self.counters[metric.name] += metric.value
def get_metrics(self, name: str = None, since: datetime = None) -> List[Metric]:
"""Get metrics by name and time range"""
with self.lock:
filtered_metrics = []
for metric in self.metrics:
# Filter by name
if name and metric.name != name:
continue
# Filter by time
if since and metric.timestamp < since:
continue
filtered_metrics.append(metric)
return filtered_metrics
def get_current_value(self, name: str) -> Optional[float]:
"""Get current value for gauge metric"""
with self.lock:
return self.current_values.get(name)
def get_counter(self, name: str) -> float:
"""Get counter value"""
with self.lock:
return self.counters.get(name, 0.0)
def get_stats(self, name: str, window_minutes: int = 60) -> Dict[str, float]:
"""Get statistics for a metric over time window"""
since = datetime.now() - timedelta(minutes=window_minutes)
metrics = self.get_metrics(name, since)
if not metrics:
return {}
values = [m.value for m in metrics]
return {
'count': len(values),
'min': min(values),
'max': max(values),
'avg': statistics.mean(values),
'median': statistics.median(values),
'p95': self._percentile(values, 95),
'p99': self._percentile(values, 99),
'latest': values[-1] if values else 0.0
}
def _percentile(self, values: List[float], percentile: int) -> float:
"""Calculate percentile"""
if not values:
return 0.0
sorted_values = sorted(values)
index = int(len(sorted_values) * percentile / 100)
return sorted_values[min(index, len(sorted_values) - 1)]
class AlertManager:
"""Manage system alerts and notifications"""
def __init__(self, storage_path: str = "logs/alerts.jsonl"):
self.storage_path = Path(storage_path)
self.storage_path.parent.mkdir(parents=True, exist_ok=True)
self.active_alerts = {}
self.alert_history = deque(maxlen=1000)
self.rules = [] # Alert rules
self.lock = threading.Lock()
def add_rule(self, name: str, metric_name: str, threshold: float,
comparison: str = "greater_than", severity: AlertSeverity = AlertSeverity.WARNING):
"""Add alert rule"""
rule = {
'name': name,
'metric_name': metric_name,
'threshold': threshold,
'comparison': comparison,
'severity': severity,
'enabled': True
}
self.rules.append(rule)
logger.info(f"Added alert rule: {name}")
def check_alerts(self, metric_collector: MetricCollector):
"""Check metrics against alert rules"""
for rule in self.rules:
if not rule['enabled']:
continue
try:
current_value = metric_collector.get_current_value(rule['metric_name'])
if current_value is None:
continue
triggered = self._evaluate_condition(
current_value, rule['threshold'], rule['comparison']
)
if triggered:
self._trigger_alert(rule, current_value, metric_collector)
else:
self._resolve_alert(rule['name'], metric_collector)
except Exception as e:
logger.error(f"Alert check failed for {rule['name']}: {e}")
def _evaluate_condition(self, value: float, threshold: float, comparison: str) -> bool:
"""Evaluate if condition is met"""
if comparison == "greater_than":
return value > threshold
elif comparison == "less_than":
return value < threshold
elif comparison == "equals":
return abs(value - threshold) < 0.001
elif comparison == "greater_equal":
return value >= threshold
elif comparison == "less_equal":
return value <= threshold
else:
return False
def _trigger_alert(self, rule: Dict[str, Any], current_value: float,
metric_collector: MetricCollector):
"""Trigger an alert"""
alert_id = rule['name']
# Check if alert is already active
if alert_id in self.active_alerts:
return
# Create new alert
alert = Alert(
id=alert_id,
severity=rule['severity'],
message=f"{rule['metric_name']} is {current_value:.2f} (threshold: {rule['threshold']})",
timestamp=datetime.now(),
metric_name=rule['metric_name'],
threshold=rule['threshold'],
current_value=current_value
)
with self.lock:
self.active_alerts[alert_id] = alert
self.alert_history.append(alert)
self._save_alert(alert)
logger.warning(f"Alert triggered: {alert.message}")
def _resolve_alert(self, alert_id: str, metric_collector: MetricCollector):
"""Resolve an active alert"""
if alert_id not in self.active_alerts:
return
with self.lock:
alert = self.active_alerts[alert_id]
alert.resolved = True
alert.resolved_at = datetime.now()
# Move to history
del self.active_alerts[alert_id]
self._save_alert(alert)
logger.info(f"Alert resolved: {alert_id}")
def _save_alert(self, alert: Alert):
"""Save alert to persistent storage"""
try:
with open(self.storage_path, 'a') as f:
alert_data = asdict(alert)
alert_data['timestamp'] = alert.timestamp.isoformat()
if alert.resolved_at:
alert_data['resolved_at'] = alert.resolved_at.isoformat()
f.write(json.dumps(alert_data) + '\n')
except Exception as e:
logger.error(f"Failed to save alert: {e}")
def get_active_alerts(self) -> List[Alert]:
"""Get currently active alerts"""
with self.lock:
return list(self.active_alerts.values())
def get_alert_history(self, limit: int = 100) -> List[Alert]:
"""Get alert history"""
with self.lock:
return list(self.alert_history)[-limit:]
class SystemMonitor:
"""Monitor system resources and health"""
def __init__(self, check_interval: int = 30):
self.check_interval = check_interval
self.running = False
self.monitor_thread = None
# System thresholds
self.thresholds = {
'cpu_usage': 80.0, # %
'memory_usage': 85.0, # %
'disk_usage': 90.0, # %
'temperature': 70.0, # Celsius
'load_average': 2.0 # per CPU core
}
def start(self, metric_collector: MetricCollector):
"""Start system monitoring"""
if self.running:
return
self.running = True
self.monitor_thread = threading.Thread(
target=self._monitor_loop,
args=(metric_collector,),
daemon=True
)
self.monitor_thread.start()
logger.info("System monitoring started")
def stop(self):
"""Stop system monitoring"""
self.running = False
if self.monitor_thread:
self.monitor_thread.join()
logger.info("System monitoring stopped")
def _monitor_loop(self, metric_collector: MetricCollector):
"""Main monitoring loop"""
while self.running:
try:
self._collect_system_metrics(metric_collector)
time.sleep(self.check_interval)
except Exception as e:
logger.error(f"System monitoring error: {e}")
time.sleep(5) # Brief pause on error
def _collect_system_metrics(self, metric_collector: MetricCollector):
"""Collect system resource metrics"""
timestamp = datetime.now()
try:
# CPU metrics
cpu_percent = psutil.cpu_percent(interval=1)
cpu_count = psutil.cpu_count()
load_avg = psutil.getloadavg()[0] if hasattr(psutil, 'getloadavg') else 0.0
metric_collector.record(Metric(
name="system.cpu.usage",
value=cpu_percent,
metric_type=MetricType.GAUGE,
timestamp=timestamp,
labels={"core": "total"}
))
metric_collector.record(Metric(
name="system.cpu.count",
value=cpu_count,
metric_type=MetricType.GAUGE,
timestamp=timestamp
))
if load_avg > 0:
metric_collector.record(Metric(
name="system.load.average",
value=load_avg,
metric_type=MetricType.GAUGE,
timestamp=timestamp
))
# Memory metrics
memory = psutil.virtual_memory()
metric_collector.record(Metric(
name="system.memory.usage",
value=memory.percent,
metric_type=MetricType.GAUGE,
timestamp=timestamp
))
metric_collector.record(Metric(
name="system.memory.available",
value=memory.available / (1024**3), # GB
metric_type=MetricType.GAUGE,
timestamp=timestamp
))
# Disk metrics
disk = psutil.disk_usage('/')
metric_collector.record(Metric(
name="system.disk.usage",
value=(disk.used / disk.total) * 100,
metric_type=MetricType.GAUGE,
timestamp=timestamp
))
# Network metrics (if available)
try:
network = psutil.net_io_counters()
metric_collector.record(Metric(
name="system.network.bytes_sent",
value=network.bytes_sent,
metric_type=MetricType.COUNTER,
timestamp=timestamp
))
metric_collector.record(Metric(
name="system.network.bytes_recv",
value=network.bytes_recv,
metric_type=MetricType.COUNTER,
timestamp=timestamp
))
except:
pass
# Process metrics
process_count = len(psutil.pids())
metric_collector.record(Metric(
name="system.processes.count",
value=process_count,
metric_type=MetricType.GAUGE,
timestamp=timestamp
))
except Exception as e:
logger.error(f"Failed to collect system metrics: {e}")
class APIMonitor:
"""Monitor API performance and usage"""
def __init__(self):
self.request_times = deque(maxlen=1000)
self.endpoint_stats = defaultdict(list)
self.error_counts = defaultdict(int)
self.lock = threading.Lock()
def record_request(self, endpoint: str, response_time: float, status_code: int):
"""Record API request metrics"""
timestamp = datetime.now()
with self.lock:
self.request_times.append({
'timestamp': timestamp,
'endpoint': endpoint,
'response_time': response_time,
'status_code': status_code
})
self.endpoint_stats[endpoint].append(response_time)
if status_code >= 400:
self.error_counts[endpoint] += 1
def get_api_stats(self, window_minutes: int = 60) -> Dict[str, Any]:
"""Get API statistics"""
since = datetime.now() - timedelta(minutes=window_minutes)
with self.lock:
recent_requests = [
req for req in self.request_times
if req['timestamp'] >= since
]
if not recent_requests:
return {}
response_times = [req['response_time'] for req in recent_requests]
error_requests = [req for req in recent_requests if req['status_code'] >= 400]
return {
'total_requests': len(recent_requests),
'error_requests': len(error_requests),
'error_rate': len(error_requests) / len(recent_requests),
'avg_response_time': statistics.mean(response_times),
'p95_response_time': self._percentile(response_times, 95),
'endpoints': {
endpoint: {
'count': len(times),
'avg_time': statistics.mean(times),
'errors': self.error_counts.get(endpoint, 0)
}
for endpoint, times in self.endpoint_stats.items()
if any(req['endpoint'] == endpoint for req in recent_requests)
}
}
def _percentile(self, values: List[float], percentile: int) -> float:
"""Calculate percentile"""
if not values:
return 0.0
sorted_values = sorted(values)
index = int(len(sorted_values) * percentile / 100)
return sorted_values[min(index, len(sorted_values) - 1)]
class MonitoringDashboard:
"""Real-time monitoring dashboard"""
def __init__(self, data_dir: str = "logs"):
self.data_dir = Path(data_dir)
self.data_dir.mkdir(exist_ok=True)
self.metric_collector = MetricCollector()
self.alert_manager = AlertManager(str(self.data_dir / "alerts.jsonl"))
self.system_monitor = SystemMonitor()
self.api_monitor = APIMonitor()
# Setup default alert rules
self._setup_default_alerts()
self.running = False
self.dashboard_thread = None
def _setup_default_alerts(self):
"""Setup default alert rules"""
# High CPU usage
self.alert_manager.add_rule(
name="high_cpu_usage",
metric_name="system.cpu.usage",
threshold=80.0,
comparison="greater_than",
severity=AlertSeverity.WARNING
)
# High memory usage
self.alert_manager.add_rule(
name="high_memory_usage",
metric_name="system.memory.usage",
threshold=85.0,
comparison="greater_than",
severity=AlertSeverity.WARNING
)
# High disk usage
self.alert_manager.add_rule(
name="high_disk_usage",
metric_name="system.disk.usage",
threshold=90.0,
comparison="greater_than",
severity=AlertSeverity.CRITICAL
)
# High API response time
self.alert_manager.add_rule(
name="high_api_response_time",
metric_name="api.response.time",
threshold=5.0,
comparison="greater_than",
severity=AlertSeverity.WARNING
)
# High error rate
self.alert_manager.add_rule(
name="high_error_rate",
metric_name="api.error.rate",
threshold=0.1, # 10%
comparison="greater_than",
severity=AlertSeverity.ERROR
)
def start(self):
"""Start monitoring dashboard"""
if self.running:
return
self.running = True
# Start system monitoring
self.system_monitor.start(self.metric_collector)
# Start dashboard update thread
self.dashboard_thread = threading.Thread(
target=self._dashboard_loop,
daemon=True
)
self.dashboard_thread.start()
logger.info("Monitoring dashboard started")
def stop(self):
"""Stop monitoring dashboard"""
self.running = False
self.system_monitor.stop()
if self.dashboard_thread:
self.dashboard_thread.join()
logger.info("Monitoring dashboard stopped")
def _dashboard_loop(self):
"""Main dashboard update loop"""
while self.running:
try:
# Update metrics
self._update_api_metrics()
# Check alerts
self.alert_manager.check_alerts(self.metric_collector)
# Save dashboard state
self._save_dashboard_state()
time.sleep(30) # Update every 30 seconds
except Exception as e:
logger.error(f"Dashboard update error: {e}")
time.sleep(10)
def _update_api_metrics(self):
"""Update API-related metrics"""
timestamp = datetime.now()
# Get API stats
api_stats = self.api_monitor.get_api_stats(window_minutes=5)
if 'avg_response_time' in api_stats:
self.metric_collector.record(Metric(
name="api.response.time",
value=api_stats['avg_response_time'],
metric_type=MetricType.GAUGE,
timestamp=timestamp
))
if 'error_rate' in api_stats:
self.metric_collector.record(Metric(
name="api.error.rate",
value=api_stats['error_rate'],
metric_type=MetricType.GAUGE,
timestamp=timestamp
))
def _save_dashboard_state(self):
"""Save current dashboard state to file"""
try:
state = {
'timestamp': datetime.now().isoformat(),
'active_alerts': [
asdict(alert) for alert in self.alert_manager.get_active_alerts()
],
'system_metrics': {
name: self.metric_collector.get_current_value(name)
for name in [
'system.cpu.usage',
'system.memory.usage',
'system.disk.usage'
]
},
'api_stats': self.api_monitor.get_api_stats()
}
# Convert datetime objects
for alert in state['active_alerts']:
alert['timestamp'] = alert['timestamp'].isoformat()
if alert['resolved_at']:
alert['resolved_at'] = alert['resolved_at'].isoformat()
state_file = self.data_dir / "dashboard_state.json"
with open(state_file, 'w') as f:
json.dump(state, f, indent=2)
except Exception as e:
logger.error(f"Failed to save dashboard state: {e}")
def record_api_request(self, endpoint: str, response_time: float, status_code: int):
"""Record API request for monitoring"""
self.api_monitor.record_request(endpoint, response_time, status_code)
def get_dashboard_data(self) -> Dict[str, Any]:
"""Get current dashboard data"""
return {
'active_alerts': [
asdict(alert) for alert in self.alert_manager.get_active_alerts()
],
'system_health': {
'cpu_usage': self.metric_collector.get_current_value('system.cpu.usage'),
'memory_usage': self.metric_collector.get_current_value('system.memory.usage'),
'disk_usage': self.metric_collector.get_current_value('system.disk.usage'),
},
'api_performance': self.api_monitor.get_api_stats(),
'recent_alerts': self.alert_manager.get_alert_history(limit=10)
}
def export_metrics(self, format: str = "json", hours: int = 24) -> str:
"""Export metrics in specified format"""
since = datetime.now() - timedelta(hours=hours)
if format.lower() == "json":
metrics_data = {
'export_timestamp': datetime.now().isoformat(),
'time_range': f"last_{hours}_hours",
'metrics': [
{
'name': metric.name,
'value': metric.value,
'timestamp': metric.timestamp.isoformat(),
'labels': metric.labels,
'type': metric.metric_type.value
}
for metric in self.metric_collector.get_metrics(since=since)
]
}
return json.dumps(metrics_data, indent=2)
else:
raise ValueError(f"Unsupported export format: {format}")
# Global dashboard instance
monitoring_dashboard = MonitoringDashboard()
# Utility functions
def test_monitoring_system():
"""Test the monitoring system"""
print("Testing monitoring system...")
dashboard = MonitoringDashboard()
# Record some test metrics
dashboard.record_api_request('/generate', 1.5, 200)
dashboard.record_api_request('/generate', 2.1, 200)
dashboard.record_api_request('/generate', 0.8, 500)
# Get dashboard data
data = dashboard.get_dashboard_data()
print(f"Active alerts: {len(data['active_alerts'])}")
print(f"API performance: {data['api_performance']}")
# Export metrics
exported = dashboard.export_metrics(format="json", hours=1)
print(f"Exported metrics: {len(exported)} characters")
print("Monitoring system test complete")
if __name__ == "__main__":
# Create logs directory
Path("logs").mkdir(exist_ok=True)
# Test monitoring functionality
test_monitoring_system()
print("\nMonitoring system ready for integration")