You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Objective
I would like to propose the addition of a feature that enables the use of a secondary Frigate installation to function as a "cold" backup system. This setup would allow users to view past detections in cases of hardware failure, theft, or other disruptions. The primary (“hot”) Frigate installation would handle all real-time processing, while the secondary system would serve as a backup viewer.
Proposed Functionality
A parameter to define the URL and authentication for a remote Frigate instance would enable these two systems (hot and cold) to integrate seamlessly. The hot system would handle all detection and processing, while the cold system would provide a view-only interface for historical data.
Current Workaround
My current workaround is as follows:
I have a script on the hot system that subscribes to MQTT topics and streams local camera feeds to a cloud Frigate instance via FFmpeg/GStreamer whenever a detection event is active.
In the cloud, Mediamtx is configured to receive these streams, with the Frigate instance set to monitor the Mediamtx paths as camera sources in short intervals. This setup ensures that the cloud Frigate only records when active detections occur.
The cloud Frigate is configured with a retention policy to store all incoming data without detection analysis (they are already performed at the edge)
If it was possible (or I knew how) to "publish on detection" and "record on incoming rtsp client" it would be much easier.
Are there simpler workarrounds that retain the ability to use the UI for historical analysis with no more than few seconds of missing footage / what is the best practice for offsite backup near realtime + low bandwidth.
Thank you
The text was updated successfully, but these errors were encountered:
Objective
I would like to propose the addition of a feature that enables the use of a secondary Frigate installation to function as a "cold" backup system. This setup would allow users to view past detections in cases of hardware failure, theft, or other disruptions. The primary (“hot”) Frigate installation would handle all real-time processing, while the secondary system would serve as a backup viewer.
Proposed Functionality
A parameter to define the URL and authentication for a remote Frigate instance would enable these two systems (hot and cold) to integrate seamlessly. The hot system would handle all detection and processing, while the cold system would provide a view-only interface for historical data.
Current Workaround
My current workaround is as follows:
If it was possible (or I knew how) to "publish on detection" and "record on incoming rtsp client" it would be much easier.
Are there simpler workarrounds that retain the ability to use the UI for historical analysis with no more than few seconds of missing footage / what is the best practice for offsite backup near realtime + low bandwidth.
Thank you
The text was updated successfully, but these errors were encountered: