Hawkeye
Detect images in a video stream and execute automated actions.
Use case
Triggering Ad insertion on an AWS MediaLive Channel when a slate image is present in the video feed.
Running locally
Running the Worker directly with Docker
The worker can be run independently with Docker or directly on the host machine.
Build Worker:
docker build -f worker.Dockerfile -t hawkeye-worker:0.0.1 .
docker run -p 5000:5000/udp -p 3030:3030 -v /home/user/dev/hawkeye/fixtures:/local -it hawkeye-worker:0.0.1 /local/watcher.jsonRunning the full Hawkeye application in Minikube
The full Hawkeye application consists of a REST API that manages the Workers using the Kubernetes API.
First we build the API docker image:
docker build -f api.Dockerfile -t hawkeye-api:0.0.1 .Prometheus metrics
The Worker expose metrics in the standard /metrics path for Prometheus to harvest.
$ curl http://localhost:3030/metrics
Environment Variables
| Environment Variable | Default | Description |
|---|---|---|
HAWKEYE_ENV |
local | dev/prod/whatever you want |
HAWKEYE_SENTRY_DSN |
the DSN url to the Sentry project to use | |
HAWKEYE_SENTRY_ENABLED |
0 |
"1" or 0 will toggle Sentry initialization |