diff --git a/secator/hooks/mongodb.py b/secator/hooks/mongodb.py
index 94e17ef11..3fd22cf66 100644
--- a/secator/hooks/mongodb.py
+++ b/secator/hooks/mongodb.py
@@ -1,4 +1,5 @@
 import logging
+import threading
 import time
 
 import pymongo
@@ -9,7 +10,7 @@
 from secator.hooks._dedup import compute_duplicate_updates
 from secator.output_types import OUTPUT_TYPES
 from secator.runners import Scan, Task, Workflow
-from secator.utils import debug, escape_mongodb_url, should_update
+from secator.utils import debug, escape_mongodb_url
 
 # import gevent.monkey
 # gevent.monkey.patch_all()
@@ -23,6 +24,10 @@
 # Time-based flushing is throttled by CONFIG.runners.backend_update_frequency.
 MONGODB_FLUSH_SIZE = 1000
 
+# Guards the per-runner findings buffer: appends happen on the main thread
+# (on_item) while flushes can run from the runner's interval thread (on_interval).
+_findings_buffer_lock = threading.Lock()
+
 logger = logging.getLogger(__name__)
 
 _mongodb_client = None
@@ -117,38 +122,50 @@ def update_finding(self, item):
 	"""
 	if type(item) not in OUTPUT_TYPES:
 		return item
+	# Add all in-memory context HERE (synchronously, before the item is yielded):
+	# the batched DB write below must never mutate the item, since the flush can
+	# run later / on another thread, after the item has already been yielded.
 	if not ObjectId.is_valid(str(item._uuid)):
 		item._uuid = str(ObjectId())
-	buffer = getattr(self, '_mongodb_findings_buffer', None)
-	if buffer is None:
-		buffer = self._mongodb_findings_buffer = []
-	buffer.append(
-		pymongo.UpdateOne({'_id': ObjectId(item._uuid)}, {'$set': item.toDict()}, upsert=True)
-	)
-	if len(buffer) >= MONGODB_FLUSH_SIZE:
+	op = pymongo.UpdateOne({'_id': ObjectId(item._uuid)}, {'$set': item.toDict()}, upsert=True)
+	with _findings_buffer_lock:
+		buffer = getattr(self, '_mongodb_findings_buffer', None)
+		if buffer is None:
+			buffer = self._mongodb_findings_buffer = []
+		buffer.append(op)
+		over_cap = len(buffer) >= MONGODB_FLUSH_SIZE
+	if over_cap:
 		flush_findings_buffer(self)
 	return item
 
 
 def flush_findings_buffer(self):
-	"""Write all buffered finding upserts to MongoDB in a single bulk_write."""
-	buffer = getattr(self, '_mongodb_findings_buffer', None)
-	if not buffer:
-		return
+	"""Write all buffered finding upserts to MongoDB in a single bulk_write.
+
+	The buffer is swapped out under a lock (so concurrent on_item appends aren't
+	lost), then written outside the lock to avoid holding it during DB I/O.
+	"""
+	with _findings_buffer_lock:
+		buffer = getattr(self, '_mongodb_findings_buffer', None)
+		if not buffer:
+			return
+		self._mongodb_findings_buffer = []
 	start_time = time.time()
 	client = get_mongodb_client()
 	db = client.main
 	count = len(buffer)
 	db.findings.bulk_write(buffer, ordered=False)
-	self._mongodb_findings_buffer = []
 	self._last_findings_flush = start_time
 	debug(f'flushed {count} findings in {time.time() - start_time:.4f}s', sub='hooks.mongodb', obj_after=False)
 
 
 def flush_findings(self):
-	"""on_interval hook: flush buffered findings, throttled by backend_update_frequency."""
-	if should_update(CONFIG.runners.backend_update_frequency, getattr(self, '_last_findings_flush', None)):
-		flush_findings_buffer(self)
+	"""on_interval hook: flush buffered findings.
+
+	Cadence is handled by run_hooks' on_interval throttle (backend_update_frequency)
+	and the runner's interval thread; this just drains the buffer when called.
+	"""
+	flush_findings_buffer(self)
 
 
 def flush_findings_final(self):
diff --git a/secator/runners/_base.py b/secator/runners/_base.py
index a6289b1d8..5cae4f775 100644
--- a/secator/runners/_base.py
+++ b/secator/runners/_base.py
@@ -3,6 +3,7 @@
 import logging
 import sys
 import textwrap
+import threading
 import uuid
 
 from datetime import datetime, timezone
@@ -107,6 +108,10 @@ def __init__(self, config, inputs=[], results=[], run_opts={}, hooks={}, validat
 		self.results = []
 		self.results_count = 0
 		self.threads = []
+		# Interval-hook thread (lazily started in __iter__ so the runner stays
+		# picklable for Celery, like the monitor thread). None until started.
+		self._interval_thread = None
+		self._interval_stop_event = None
 		self.output = ''
 		self.started = False
 		self.done = False
@@ -458,6 +463,9 @@ def __getstate__(self):
 		state = self.__dict__.copy()
 		state['_hooks'] = {}
 		state['resolved_hooks'] = {name: [] for name in state['resolved_hooks']}
+		# Threads/Events aren't picklable; they're recreated lazily on run.
+		state['_interval_thread'] = None
+		state['_interval_stop_event'] = None
 		return state
 
 	def __setstate__(self, state):
@@ -545,6 +553,9 @@ def __iter__(self):
 			else:
 				self.log_start()
 
+			# Fire on_interval hooks on a timer (covers quiet periods); stopped in _finalize
+			self._start_interval_thread()
+
 			# Yield results buffer
 			yield from self.results_buffer
 			self.results_buffer = []
@@ -577,6 +588,9 @@ def __iter__(self):
 
 	def _finalize(self):
 		"""Finalize the runner."""
+		# Stop interval thread first so it can't flush concurrently with the
+		# final on_end flush (mark_completed below).
+		self._stop_interval_thread()
 		self.join_threads()
 		gc.collect()
 		if self.sync:
@@ -595,6 +609,38 @@ def _finalize(self):
 		if self.enable_reports:
 			self.export_reports()
 
+	def _start_interval_thread(self):
+		"""Start a daemon thread firing on_interval hooks on a time interval.
+
+		Created here (not in __init__) so the runner stays picklable for Celery,
+		mirroring the monitor thread. Without this, on_interval only fires when the
+		runner produces an item, so it stalls during quiet periods. on_interval is
+		still throttled by backend_update_frequency in run_hooks; a value <= 0
+		disables time-based backend updates, so we don't start the thread at all.
+		"""
+		frequency = CONFIG.runners.backend_update_frequency
+		if frequency <= 0 or self._interval_thread is not None:
+			return
+		self._interval_stop_event = threading.Event()
+		self._interval_thread = threading.Thread(target=self._interval_loop, args=(frequency,), daemon=True)
+		self._interval_thread.start()
+
+	def _interval_loop(self, frequency):
+		"""Fire on_interval hooks every `frequency` seconds until stopped."""
+		while not self._interval_stop_event.wait(frequency):
+			try:
+				self.run_hooks('on_interval', sub='interval')
+			except Exception as e:
+				self.debug(f'interval thread hook error: {e}', sub='interval')
+
+	def _stop_interval_thread(self):
+		"""Stop the interval thread (called before the final on_end flush)."""
+		if self._interval_thread and self._interval_stop_event:
+			self._interval_stop_event.set()
+			self._interval_thread.join(timeout=2.0)
+		self._interval_thread = None
+		self._interval_stop_event = None
+
 	def join_threads(self):
 		"""Wait for all running threads to complete."""
 		if not self.threads:
@@ -938,9 +984,11 @@ def toDict(self):
 				'output': self.output,
 				'progress': self.progress,
 				'last_updated_db': self.last_updated_db,
-				'context': {**self.context, 'celery_ids': list(self.celery_ids_map.keys())},
-				'errors': [e.toDict() for e in self.errors],
-				'warnings': [w.toDict() for w in self.warnings],
+				# Snapshot mutable collections (copy before iterating): toDict can be
+				# called from the interval thread while the main thread appends.
+				'context': {**self.context, 'celery_ids': list(self.celery_ids_map.copy())},
+				'errors': [e.toDict() for e in list(self.errors)],
+				'warnings': [w.toDict() for w in list(self.warnings)],
 			}
 		)
 		return data