The Watson Machine Learning Quickstart demonstrates use of PostgreSQL, Watson Machine Learning, and MongoDB on Cloud Pak for Data to continuously enrich enterprise data with machine learning insights. This quickstart will get you up and running on any OpenShift cluster including a local minishift running on your machine.
In this example, we have one microservice producing simulated refrigeration container (reefer) telemetry events that include temperature, humidity, cumulative power consumption, etc. The events are persisted in a PostgreSQL database. A second microservice polls the telemetry data, applies a machine learning model to determine whether or not a given refrigeration unit requires maintenance, and stores the results in a MongoDB collection. The most recent results are displayed in a simple web application.
Make your data ready for an AI and multicloud world. Cloud Pak for Data System is an all-in-one cloud-native Data and AI platform in a box, providing a pre-configured, governed, and secure environment to collect, organize and analyze data. Learn more.
Installing Cloud Pak for Data on OpenShift- instructions
Installing the Watson Machine Learning add-on - instructions
Installing the PostgreSQL add-on instructions see the PostgreSQL section.
Installing the MongoDB add-on instructions see the MongoDB section.
Fork a copy of this repo
After logging in with oc login, ensure that you have a project set up. If not, create one as follows:
$ oc new-project watson-machine-learning-project --display-name="Watson Machine Learning Project"
That's it, project has been created. Ensure that your current project is set:
$ oc project watson-machine-learning-project
The template for this example is located at cpd-quick-start-watson-machine-learning.json.
First, list the available parameters:
$ oc process --parameters -f https://raw.githubusercontent.com/estherhi/cpd-quick-start-watson-machine-learning/master/openshift/templates/cpd-quick-start-watson-machine-learning.json
The following parameters are required:
- POSTGRESQL_HOST
- POSTGRESQL_USER
- POSTGRESQL_PASSWORD
- POSTGRESQL_DATABASE_NAME
- MONGODB_HOST
- MONGODB_USER
- MONGODB_PASSWORD
- MONGODB_DATABASE
- ICP4D_CLUSTER_HOST
- ICP4D_CLUSTER_USER
- ICP4D_CLUSTER_PASSWORD
Create the app from the template and specify the source url to be your forked repo:
$ oc new-app -f \
https://raw.githubusercontent.com/estherhi/cpd-quick-start-watson-machine-learning/master/openshift/templates/cpd-quick-start-watson-machine-learning.json \
-p POSTGRESQL_HOST=<POSTGRESQL_HOST> \
-p POSTGRESQL_USER=<POSTGRESQL_USER> \
-p POSTGRESQL_PASSWORD=<POSTGRESQL_PASSWORD> \
-p POSTGRESQL_DATABASE=<POSTGRESQL_DATABASE_NAME>\
-p MONGODB_HOST=<MONGO_HOST> \
-p MONGODB_USER=<MONGO_USER> \
-p MONGODB_PASSWORD=<MONGO_PASSWORD> \
-p MONGODB_DATABASE=<MONGO_DATABASE_NAME> \
-p ICP4D_CLUSTER_HOST=<ICP4D_CLUSTER_HOST> \
-p ICP4D_CLUSTER_USER=<ICP4D_CLUSTER_USER> \
-p ICP4D_CLUSTER_PASSWORD=<ICP4D_CLUSTER_PASSWORD>
oc new-app will kick off a build once all required dependencies are confirmed.
Check the status of your new nodejs app with the command:
$ oc status
Which should return something like:
In project Watson Machine Learning Project (watson-machine-learning-project) on server https://10.2.2.2:8443
svc/watson-assistant-quickstart - 172.30.108.183:8080
dc/watson-assistant-quickstart deploys istag/watson-assistant-quickstart:latest <-
bc/watson-assistant-quickstart source builds https://github.ibm.com/icp4d-devex-prototype/cpd-quickstart-watson-assistant on openshift/nodejs:10
build #1 running for 7 seconds
deployment #1 waiting on image or update
Which should return something like:
In project e2e-demo on server https://192.168.42.218:8443
http://watson-machine-learning-event-scorer-e2e-demo.192.168.42.218.nip.io (svc/watson-machine-learning-event-scorer)
dc/watson-machine-learning-event-scorer deploys istag/watson-machine-learning-event-scorer:latest <-
bc/watson-machine-learning-event-scorer source builds https://github.com/estherhi/cpd-quick-start-watson-machine-learning on openshift/python:3.6
build #1 running for 59 seconds - 14e1b33: iml git ignore (estherh <estherh@il.ibmcom>)
deployment #1 waiting on image or update
dc/container-event-producer deploys istag/container-event-producer:latest <-
bc/container-event-producer source builds https://github.com/estherhi/cpd-quick-start-watson-machine-learning on openshift/python:3.6
build #1 running for 59 seconds - 14e1b33: iml git ignore (estherh <estherh@il.ibmcom>)
deployment #1 waiting on image or update
An OpenShift route exposes a service at a host name, like www.example.com, so that external clients can reach it by name.
DNS resolution for a host name is handled separately from routing; you may wish to configure a cloud domain that will always correctly resolve to the OpenShift router, or if using an unrelated host name you may need to modify its DNS records independently to resolve to the router.
That aside, let's explore our new web app. oc new-app created a new route. To view your new route:
$ oc get route
In the result you can find all routes in your project and for each route you can find its hostname.
Find the watson-assistant-quickstart route and use the hostname to navigate to the newly created Node.js web app.
Notice that you can use the APPLICATION_DOMAIN template parameter to define a hostname for your app.
To create a new route at a host name, like www.example.com:
$ oc expose svc/watson-assistant-quickstart --hostname=www.example.com
If the build is not yet started (you can check by running oc get builds), start one and stream the logs with:
$ oc start-build watson-assistant-quickstart --follow
Deployment happens automatically once the new application image is available. To monitor its status either watch the web console or execute oc get pods to see when the pod is up. Another helpful command is
$ oc get svc
This will help indicate what IP address the service is running, the default port for it to deploy at is 8080. Output should look like:
NAME CLUSTER-IP EXTERNAL-IP PORT(S) AGE
watson-assistant-quickstart 172.30.249.251 <none> 8080/TCP 7m
Assuming you used the URL of your own forked repository, you can configure your github repository to make a webhook call whenever you push your code. Learn more about Webhook Triggers.
- From the OpenShift web console homepage, navigate to your project
- Go to Builds
- Click the link with your BuildConfig name
- Click the Configuration tab
- Click the "Copy to clipboard" icon to the right of the "GitHub webhook URL" field
- Navigate to your repository on GitHub and click on repository settings > webhooks > Add webhook
- Paste your webhook URL provided by OpenShift
- Leave the defaults for the remaining fields - That's it!
- After you save your webhook, refresh your Github settings page and check the status to verify connectivity.
Learn more about OpenShift templates.
- Model versions supported in Watson Machine Learning are documented here - https://docs-icpdata.mybluemix.net/docs/content/SSQNUZ_current/com.ibm.icpdata.doc/dsx/models.html.
- Each time the example is run a model is stored in Watson Machine Learning instance and a new deployment is generated.
- The micro services are currently configured to accept insecure endpoints for Cloud Pak for Data. For production, use secure endpoints only.
This code pattern is licensed under the Apache Software License, Version 2. Separate third party code objects invoked within this code pattern are licensed by their respective providers pursuant to their own separate licenses. Contributions are subject to the Developer Certificate of Origin, Version 1.1 (DCO) and the Apache Software License, Version 2.