Bump to v1.5.0: deduplicate detectors, fix aggregator bugs, fix blocking I/O

- Extract shared send_event/clear_event into detectors/base.py, removing
  ~150 lines of duplication across all 6 detectors
- Fix default aggregator URL from port 5000 to 5100 in all detectors
- Standardize cpu.py and memory.py to use active_alerts set pattern
- Fix immediate emote rotation on startup (last_emote_change = time.time())
- Extract magic numbers to named constants in aggregator
- Protect write_status() with try/except OSError
- Fix notify event ID collision with monotonic counter
- Replace blocking stream_output() with background daemon threads in kao.py

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

detectors/service.py

@@ -3,7 +3,7 @@ Service Health Detector
 Monitors if specific processes/services are running.
 
 Environment variables:
-    AGGREGATOR_URL  - URL of the aggregator (default: http://localhost:5000)
+    AGGREGATOR_URL  - URL of the aggregator (default: http://localhost:5100)
     CHECK_INTERVAL  - Seconds between checks (default: 30)
     SERVICES        - Comma-separated list of process names to monitor (required)
                       Example: "nginx,postgres,redis"
@@ -13,10 +13,11 @@ import os
 import sys
 import time
 import psutil
-import requests
+
+from detectors.base import DEFAULT_AGGREGATOR_URL, send_event, clear_event
 
 # Configuration from environment
-AGGREGATOR_URL = os.environ.get("AGGREGATOR_URL", "http://localhost:5000")
+AGGREGATOR_URL = os.environ.get("AGGREGATOR_URL", DEFAULT_AGGREGATOR_URL)
 CHECK_INTERVAL = int(os.environ.get("CHECK_INTERVAL", 30))
 SERVICES = os.environ.get("SERVICES", "")
 
@@ -37,34 +38,6 @@ def get_running_processes():
     return running
 
 
-def send_event(event_id, priority, message):
-    """Send an event to the aggregator with heartbeat TTL."""
-    ttl = CHECK_INTERVAL * 2
-    try:
-        response = requests.post(
-            f"{AGGREGATOR_URL}/event",
-            json={"id": event_id, "priority": priority, "message": message, "ttl": ttl},
-            timeout=5
-        )
-        print(f"[EVENT] {event_id}: {message} (priority {priority}, ttl {ttl}s) -> {response.status_code}")
-    except requests.RequestException as e:
-        print(f"[ERROR] Failed to send event: {e}")
-
-
-def clear_event(event_id):
-    """Clear the event from the aggregator."""
-    try:
-        response = requests.post(
-            f"{AGGREGATOR_URL}/clear",
-            json={"id": event_id},
-            timeout=5
-        )
-        if response.status_code == 200:
-            print(f"[CLEAR] {event_id}")
-    except requests.RequestException as e:
-        print(f"[ERROR] Failed to clear event: {e}")
-
-
 def main():
     if not SERVICES:
         print("ERROR: SERVICES environment variable is required")
@@ -90,14 +63,14 @@ def main():
             event_id = f"service_{service}"
 
             if service not in running:
-                send_event(event_id, 1, f"Service '{service}' is not running")
+                send_event(AGGREGATOR_URL, event_id, 1, f"Service '{service}' is not running", CHECK_INTERVAL)
                 current_alerts.add(event_id)
             else:
                 print(f"[OK] Service '{service}' is running")
 
         # Clear alerts for services that are now running
         for event_id in active_alerts - current_alerts:
-            clear_event(event_id)
+            clear_event(AGGREGATOR_URL, event_id)
 
         active_alerts = current_alerts