adsist-cms/python-rtsp-worker
3
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

expo backoff decode handling

Browse source
This commit is contained in:
Siwat Sirichai 2025-11-11 02:30:17 +07:00
parent 1432eb4b97
commit 42b08c6251
1 changed files with 17 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

27
services/stream_decoder.py
View file

@ -380,23 +380,30 @@ class StreamDecoder:
# Set the CUDA device for this thread # Set the CUDA device for this thread
torch.cuda.set_device(self.gpu_id) torch.cuda.set_device(self.gpu_id)
retry_count = 0 consecutive_failures = 0
max_retries = 5 backoff_delay = 2.0 # Start with 2 seconds
max_backoff = 30.0 # Cap at 30 seconds
while not self._stop_flag.is_set(): while not self._stop_flag.is_set():
# Initialize connection # Initialize connection
if not self._init_rtsp_connection(): if not self._init_rtsp_connection():
retry_count += 1 consecutive_failures += 1
if retry_count >= max_retries:
print(f"Max retries reached for {self.rtsp_url}")
self._set_status(ConnectionStatus.ERROR)
break
# Exponential backoff: 2s, 4s, 8s, 16s, 30s, 30s, ...
backoff_delay = min(
2.0 * (2 ** (consecutive_failures - 1)), max_backoff
)
print(
f"Connection failed for {self.rtsp_url}, retry in {backoff_delay:.0f}s (attempt {consecutive_failures})"
)
self._set_status(ConnectionStatus.RECONNECTING) self._set_status(ConnectionStatus.RECONNECTING)
self._stop_flag.wait(timeout=2.0) self._stop_flag.wait(timeout=backoff_delay)
continue continue
retry_count = 0 # Reset on successful connection # Successfully connected - reset failure tracking
consecutive_failures = 0
backoff_delay = 2.0
try: try:
# Decode loop - iterate through packets from PyAV # Decode loop - iterate through packets from PyAV
@ -473,7 +480,7 @@ class StreamDecoder:
print(f"Error in decode loop for {self.rtsp_url}: {e}") print(f"Error in decode loop for {self.rtsp_url}: {e}")
self._set_status(ConnectionStatus.RECONNECTING) self._set_status(ConnectionStatus.RECONNECTING)
self._cleanup() self._cleanup()
self._stop_flag.wait(timeout=2.0) # Will retry connection at top of loop with exponential backoff
def _cleanup(self): def _cleanup(self):
"""Cleanup resources""" """Cleanup resources"""