Ok v4.1.0-alpha3 now stores plates locally. i'm still testing it, if you run the agent scrape it should pull in all the images since the last time it was run, right now old images won't show up. I also added a db index that should make scraping run a little faster.
you can make the rekor agent store lots of images in the settings:
store_plates_maxsize_gb = 4000
store_plates_maxsize_mb = 4000000
Rekor changed the configuration key in the latest version so i have both
I thought about having the processor store the images on its own so the agent could be deleted and recreated without affecting the images. is that what you mean?
It now only populates the make/model etc if the confidence is higher than 30%. Glad you like the changes. if you just use latest in docker it will automatically update whenever a new version gets released.
Just copy the database files to the new location and it should pick them up, it's the same for both windows and docker
EDIT: oh you're talking about the openalpr agent, i'm not sure you should probably just be able to copy the whole folder
Do you have your agent exposing the web ui port? can you get to the webpage from the machine that is running the processor?
Also make sure your url is configured correctly in the processor settings
v4.0.0-alpha2 has
1. support for filtering for plates only seen 10 times or less.
2. upgrading to dotnet 6.0
It's just a preview for now, back up your database before trying it out if you want to test the new feature.
If you are running the windows executable (not docker) you will need to...
You might just have a docker issue with letting the container talk to the outside network. I'm using this on unraid so it's all configured for me. you can always try getting inside the container with shell and pinging stuff on your local network to see if it can see anything.
Or how many cars are you seeing per minute? I believe most of the current users only see a few a minute, I haven't stress tested this against multiple cars a second if you live on a busy road.
some of your webhooks are taking around 10 seconds to process
Executed action "OpenAlprWebhookProcessor.WebhookProcessor.WebhookController.Post (OpenAlprWebhookProcessor)" in 7523.514ms
It looks like the cloud isn't liking how long it is taking the processor to execute, and is timing out
--->...
if you want something more robust i made a tool that listens for those same webhooks and gives you a nice UI to view, complete with customizable alerts that can be sent via pushover also. OpenALPR Webhook Processor for IP Cameras
In the debug logs you pasted it looks like the rekor cloud is sending you a JSON structure the processor can't handle.
It has some extra information at the beginning of the payload:
https://support.rekor.ai/my/s/article/How-to-receive-data-with-WebHooks?language=en_US says it should have a...
ya going back to previous versions won't always work if a database migration has taken place during a release. narrowing it down between 3.7 and 3.8 helps though thanks :)
Unfortunately I don't have time to run through each release version to check to see when it failed. if you start with version 12, see if it happens or not, then do release 7 or 14 depending on the outcome etc, it should be pretty fast.
Try out v3.17.0-alpha6
Go to the agent settings page and enable debug mode.
Wait for a failure to happen.
Go to the logs page and download the failed logs
paste them in here and i'll see what i can do