modules/server_monitor.py

import redis
import os
import time
from datetime import datetime, timedelta

class ServerMonitor:
    def __init__(self):
        try:
            self.redis_client = redis.Redis(
                host=os.getenv("REDIS_HOST", "localhost"),
                port=int(os.getenv("REDIS_PORT", 6379)),
                username=os.getenv("REDIS_USERNAME"),
                password=os.getenv("REDIS_PASSWORD"),
                decode_responses=True
            )
            # Test connection
            self.redis_client.ping()
            self.connected = True
        except Exception:
            self.redis_client = None
            self.connected = False
    
    def report_failure(self):
        """Report a server failure (e.g., 503 error)"""
        if not self.connected:
            return
            
        try:
            # Increment failure counter
            key = f"server_failures:{datetime.now().strftime('%Y-%m-%d:%H')}"
            self.redis_client.incr(key)
            self.redis_client.expire(key, 3600)  # Expire in 1 hour
            
            # Record last failure time
            self.redis_client.set("last_failure", datetime.now().isoformat())
            self.redis_client.expire("last_failure", 86400)  # Expire in 24 hours
        except Exception:
            pass  # Silently fail to avoid breaking the main app
    
    def report_success(self):
        """Report a successful request"""
        if not self.connected:
            return
            
        try:
            # Reset failure counter for current hour
            key = f"server_failures:{datetime.now().strftime('%Y-%m-%d:%H')}"
            self.redis_client.delete(key)
            
            # Record last success time
            self.redis_client.set("last_success", datetime.now().isoformat())
            self.redis_client.expire("last_success", 86400)  # Expire in 24 hours
        except Exception:
            pass  # Silently fail to avoid breaking the main app
    
    def check_server_status(self):
        """Check if server is likely available based on recent activity"""
        if not self.connected:
            return {"available": True, "message": "Redis not configured, assuming server available"}
        
        try:
            # Get recent failures
            now = datetime.now()
            failures_last_hour = 0
            
            # Check current and previous hour
            for i in range(2):
                check_time = now - timedelta(hours=i)
                key = f"server_failures:{check_time.strftime('%Y-%m-%d:%H')}"
                failures = self.redis_client.get(key)
                if failures:
                    failures_last_hour += int(failures)
            
            # Get last failure time
            last_failure_str = self.redis_client.get("last_failure")
            last_success_str = self.redis_client.get("last_success")
            
            # If we had recent failures but no recent success, server might be down
            if failures_last_hour > 3:
                if last_success_str:
                    last_success = datetime.fromisoformat(last_success_str)
                    minutes_since_success = (now - last_success).total_seconds() / 60
                    if minutes_since_success < 15:
                        return {
                            "available": True, 
                            "message": "Recent success detected, server likely available",
                            "estimated_wait": 0
                        }
                
                # Estimate wait time based on typical warmup
                return {
                    "available": False, 
                    "message": f"High failure rate detected ({failures_last_hour} failures recently)",
                    "estimated_wait": 5
                }
            
            # If we had a very recent failure (< 5 mins), suggest waiting
            if last_failure_str:
                last_failure = datetime.fromisoformat(last_failure_str)
                minutes_since_failure = (now - last_failure).total_seconds() / 60
                if minutes_since_failure < 5:
                    return {
                        "available": False, 
                        "message": f"Recent failure {int(minutes_since_failure)} minutes ago",
                        "estimated_wait": max(1, 5 - int(minutes_since_failure))
                    }
            
            return {
                "available": True, 
                "message": "Server appears to be available",
                "estimated_wait": 0
            }
            
        except Exception as e:
            # On any Redis error, assume server is available
            return {
                "available": True, 
                "message": f"Monitoring check failed: {str(e)}, assuming server available",
                "estimated_wait": 0
            }
    
    def get_system_stats(self):
        """Get detailed system statistics"""
        if not self.connected:
            return {"error": "Redis not configured"}
        
        try:
            stats = {}
            
            # Get recent failures
            now = datetime.now()
            total_failures = 0
            for i in range(24):  # Last 24 hours
                check_time = now - timedelta(hours=i)
                key = f"server_failures:{check_time.strftime('%Y-%m-%d:%H')}"
                failures = self.redis_client.get(key)
                if failures:
                    total_failures += int(failures)
            
            stats["failures_last_24h"] = total_failures
            
            # Get last events
            last_failure = self.redis_client.get("last_failure")
            last_success = self.redis_client.get("last_success")
            
            stats["last_failure"] = last_failure if last_failure else "None recorded"
            stats["last_success"] = last_success if last_success else "None recorded"
            
            # Calculate uptime percentage (approximate)
            if last_failure and last_success:
                failure_time = datetime.fromisoformat(last_failure)
                success_time = datetime.fromisoformat(last_success)
                if success_time > failure_time:
                    stats["status"] = "Operational"
                else:
                    stats["status"] = "Degraded"
            elif last_success:
                stats["status"] = "Operational"
            elif last_failure:
                stats["status"] = "Issues Detected"
            else:
                stats["status"] = "Unknown"
                
            return stats
            
        except Exception as e:
            return {"error": str(e)}