Use _stop_event consistently across all threads as the way to signal that it should stop. Add thread joins with timeouts to allow the program to exit cleanly

core/cleanup.py

@@ -34,6 +34,10 @@ class CleanupTimer:
         """Stop any threads and prepare for application shutdown"""
 
         self._stop_event.set()
+        if self._thread:
+            self._thread.join(timeout=15)
+            if self._thread.is_alive():
+                logger.warning("Cleanup worker thread did not exit on time and will be killed.")
 
     def _run(self):
         while not self._stop_event.wait(timeout=self._cleanup_interval):

core/data_providers.py

@@ -1,5 +1,6 @@
 import logging
 import threading
+import time
 
 from core.config import config, create_provider_from_config
 
@@ -16,6 +17,7 @@ class DataProviders:
         self.static_data_providers = []
         self.sig_ref_data_providers = []
         self.callsign_data_providers = []
+        self._startup_timers = []
 
     def setup(self):
         for entry in config["spot_providers"]:
@@ -43,41 +45,58 @@ class DataProviders:
     def start(self):
         # Start data providers before spot/alert providers so the lookup data is there already for incoming spots.
         # Each category is fired off after a small delay to give the rest of Spothole chance to start up.
-        threading.Timer(5.0, lambda: self.start_providers(self.static_data_providers, "static data")).start()
-        threading.Timer(
-            10.0,
-            lambda: self.start_providers(self.callsign_data_providers, "callsign data"),
-        ).start()
-        threading.Timer(15.0, lambda: self.start_providers(self.spot_providers, "spot")).start()
-        threading.Timer(20.0, lambda: self.start_providers(self.alert_providers, "alert")).start()
-        threading.Timer(
-            25.0,
-            lambda: self.start_providers(self.solar_condition_providers, "solar condition"),
-        ).start()
-        threading.Timer(
-            30.0,
-            lambda: self.start_providers(self.sig_ref_data_providers, "SIG ref data"),
-        ).start()
+        self._startup_timers = [
+            threading.Timer(5.0, lambda: self.start_providers(self.static_data_providers, "static data")),
+            threading.Timer(10.0, lambda: self.start_providers(self.callsign_data_providers, "callsign data")),
+            threading.Timer(15.0, lambda: self.start_providers(self.spot_providers, "spot")),
+            threading.Timer(20.0, lambda: self.start_providers(self.alert_providers, "alert")),
+            threading.Timer(
+                25.0,
+                lambda: self.start_providers(self.solar_condition_providers, "solar condition"),
+            ),
+            threading.Timer(30.0, lambda: self.start_providers(self.sig_ref_data_providers, "SIG ref data")),
+        ]
+        for t in self._startup_timers:
+            t.daemon = True
+            t.start()
 
     def stop(self):
-        for sp in self.spot_providers:
-            if sp.enabled:
-                sp.stop()
-        for ap in self.alert_providers:
-            if ap.enabled:
-                ap.stop()
-        for scp in self.solar_condition_providers:
-            if scp.enabled:
-                scp.stop()
-        for srdp in self.sig_ref_data_providers:
-            if srdp.enabled:
-                srdp.stop()
-        for sdp in self.static_data_providers:
-            if sdp.enabled:
-                sdp.stop()
-        for cdp in self.callsign_data_providers:
-            if cdp.enabled:
-                cdp.stop()
+        # Cancel any startup timers that haven't fired yet
+        for t in self._startup_timers:
+            t.cancel()
+
+        # Stop all providers
+        all_providers = [
+            p
+            for p in (
+                self.spot_providers
+                + self.alert_providers
+                + self.solar_condition_providers
+                + self.sig_ref_data_providers
+                + self.static_data_providers
+                + self.callsign_data_providers
+            )
+            if p.enabled
+        ]
+        if not all_providers:
+            return
+
+        def stop_provider(p):
+            try:
+                p.stop()
+            except Exception:
+                logger.exception("Exception stopping provider")
+
+        threads = [threading.Thread(target=stop_provider, args=(p,), daemon=True) for p in all_providers]
+        for t in threads:
+            t.start()
+
+        deadline = time.monotonic() + 40
+        for t in threads:
+            t.join(timeout=max(0.0, deadline - time.monotonic()))
+        still_running = [t for t in threads if t.is_alive()]
+        if still_running:
+            logger.warning("Some threads did not stop in time!")
 
 
 # Global object

core/live_data_cache.py

@@ -23,6 +23,7 @@ class LiveDataCache:
         self._snapshot_dir = snapshot_dir
         self._disk_cache = diskcache.Cache(str(snapshot_dir))
         self._stop_event = threading.Event()
+        self._snapshot_thread = None
         self._load_snapshot()
         self._start_periodic_snapshot(snapshot_interval_sec)
 
@@ -93,10 +94,16 @@ class LiveDataCache:
             while not self._stop_event.wait(timeout=interval):
                 self.save_snapshot()
 
-        t = threading.Thread(target=loop, name=f"LiveDataCache-Snapshot-{self._snapshot_dir}", daemon=True)
-        t.start()
+        self._snapshot_thread = threading.Thread(
+            target=loop, name=f"LiveDataCache-Snapshot-{self._snapshot_dir}", daemon=True
+        )
+        self._snapshot_thread.start()
 
     def close(self):
         self._stop_event.set()
+        if self._snapshot_thread:
+            self._snapshot_thread.join(timeout=15)
+            if self._snapshot_thread.is_alive():
+                logger.warning(f"LiveDataCache snapshot thread for {self._snapshot_dir} did not exit on time.")
         self.save_snapshot()
         self._disk_cache.close()

core/status_reporter.py

@@ -1,3 +1,4 @@
+import logging
 import os
 from datetime import datetime
 from threading import Event, Thread
@@ -14,6 +15,8 @@ from core.prometheus_metrics_handler import alerts_gauge, memory_use_gauge, spot
 from telnetserver.telnetserver import TELNET_SERVER
 from webserver.webserver import WEB_SERVER
 
+logger = logging.getLogger(__name__)
+
 
 class StatusReporter:
     """Provides a timed update of the application's status data."""
@@ -33,13 +36,17 @@ class StatusReporter:
     def start(self):
         """Start the reporter thread"""
 
-        self._thread = Thread(target=self._run, name="StatusReporter")
+        self._thread = Thread(target=self._run, name="StatusReporter", daemon=True)
         self._thread.start()
 
     def stop(self):
         """Stop any threads and prepare for application shutdown"""
 
         self._stop_event.set()
+        if self._thread:
+            self._thread.join(timeout=15)
+            if self._thread.is_alive():
+                logger.warning("Status reporter worker thread did not exit on time and will be killed.")
 
     def _run(self):
         """Thread entry point: report immediately on startup, then on each interval until stopped"""