Dog Detector Sample App Setup

A Sample App that detects dogs

• Kafka instance
• Web application
• Dog detector REST service
• Dog detector Kafka consumer

Prerequesites:

• OpenShift (tested on 4.7)
• Installed Strimzi Operator (tested on 0.23.0)

Deploying to an OpenShift Cluster

Login Information

Customize the .env.local file to include OpenShift login information

# using token
OC_URL=https://api.cluster:6443
OC_TOKEN=sha256~blahblah
OC_PROJECT=dog-detector

or

# using username & password
OC_URL=https://api.cluster:6443
OC_USER=your-username
OC_PASSWORD=your-password
OC_PROJECT=dog-detector

Test with:

$ make login

Using your Fork

If you forked the service/consumer repos, you can edit the .env.local to build from your own source to image repository.

REST_SERVICE_GIT_REPO=https://github.com/your-org/dog-detector-service.git
KAFKA_CONSUMER_GIT_REPO=https://github.com/your-org/dog-detector-kafka-consumer.git

Execute Deployment

Now execute the deployment scripts.

$ make deploy

Navigate to the Application

Navigate to the URL in the route dog-detector-app. To find it, you can query:

echo "https://$(oc get route dog-detector-app -o jsonpath='{.spec.host}')"

Known Issues

• The first request to Tensorflow is slow as it loads up the model.