If this helps, I have it working but I don’t have a whitelist setup but this limits you to where you can have folder watcher monitor files.
For me under homeassistant’s config directory www/ I created subdirectories to monitor.
I get the same thing - doesn’t work for me. No errors logged.
(the sub folder under www is a link to a smb share - but, access is fine - motion, etc is happily writing there, I can also access items such as jpg files from home-assistant on the share)
The watched folder needs to be accessible directly by the OS, and will not work on network drives. Please see the docs to understand how watchdog operates https://pythonhosted.org/watchdog/
I use Ubuntu and I locally mounted a samba (cifs) to a remote file system.
I also added the newly mounted directory to the white list. Folder_sensor worked.
It took anywhere from a couple of seconds to as much as 35 seconds to recognize
file adds/deletes.
Think I might look into using something else. My requirement is quite simple, I just want to send a telegram video message when motion is captured on an outdoor camera.
I would have been nice to do as an automation for the ability to easily disable or enable the “service”
Hello, did you get an answer for this after? I’m in the same situation, currently sending a notification when my camera detects a person, but sometimes the media isn’t available straight away (thanks nest) so want to trigger based on a file being created instead of the device
I think i’ve figured this out now. Its over engineered, but seems to be working more consistently, and get access to a higher image in the notification (the nest ones were blurry)
So it uses Watch Folder to check when a new file is created in the specified folder, that will trigger my shell command that will simply copy the newest file from the nest directory , and copy it to the www folder at ‘latest’ (Used copy to keep the originals in nest folder incase I ever need them & move to latest in destination folder so that this folder doesn’t fill up)
The notification simply uses the latest.jpg image each time, which so far get’s updated and haven’t noticed a significant delay from the original way of doing it (both automations are running while i assess) and of course, each time you have a valid image