Installing Process Mining in a Cloud Pak

You can use IBM Process Mining within one of the following IBM Cloud Pak:

  • IBM Cloud Pak for Business Automation.
  • IBM Cloud Pak for Integration.
  • IBM Cloud Pak for Network Automation.
  • {{site.data.keyword.cp4waiops_notm}}.

Installing IBM Process Mining requires the additional involvement of a dedicated operator managing the installation. To use IBM Process Mining in a Cloud Pak:

  • Install the Cloud Pak. For more information about the installation of Cloud Paks, see IBM Cloud Paks documentation.

  • Install the Process Mining capabilities in the Cloud Pak user interface. Use the IBM Process Mining (PM) operator which defines a Process Mining Custom Resource Definition (CRD) in Kubernetes.

  • Deploy the IBM Process Mining CRD in the same Kubernetes namespace. IBM Process Mining augments the Cloud Pak user interface with new menu entries and user roles.

  • Specify the name of the Cloud Pak in the CRD during provisioning of the instance, so the metering of your IBM Process Mining usage corresponds to the Cloud Pak you are entitled to. The Cloud Pak selection list is available in the Form View during instance provisioning.

You can deploy the IBM Process Mining CRD in a namespace that does not have any Cloud Pak installed.

You can install IBM Process Mining in several namespaces in the same cluster, as all the resources are contained in the same namespace.

You need to configure IBM Task Mining to use an external database if the the Cloud Pak already has a Db2 database instance with IBM Cloud Pak for Business Automation.

Installing PM without TM

The following configuration example shows the possibility to install IBM Process Mining without IBM Task Mining and Db2 configurations.

apiVersion: processmining.ibm.com/v1beta1
kind: ProcessMining
metadata:
  name: pm
  namespace: processmining
spec:
  version : 1.15.0
  license:
    accept: true
    cloudPak: IBM Cloud Pak for Business Automation
  defaultStorageClassName: managed-nfs-storage
  taskmining: 
    install: false

Installing PM with a pre-existing instance of Db2

You have to adopt an existing Db2 instance, because it is impossible to install an embedded one. To proceed, get the connection information from the Db2 instance provided by IBM Cloud Pak for Business Automation. Use oc get db2ucluster to get the name of the running Db2 instance.

You can find the required information in the following list:

  • Secret name where user password is stored The name is c-<DB2 INSTANCE NAME>-instancepassword.
  • Secret key where user password is stored In IBM Cloud Pak for Business Automation default value is password.
  • Db2 hostname c-<DB2 INSTANCE NAME>-db2u.<NAMESPACE>.svc
  • Database name In IBM Cloud Pak for Business Automation default value is CP4ADB.
  • Database username In IBM Cloud Pak for Business Automation default value is db2inst1.
  • Database schema It is the user-defined name, for example, TMUSER.
apiVersion: processmining.ibm.com/v1beta1
kind: ProcessMining
metadata:
  name: pm
  namespace: processmining
spec:
  version : 1.15.0
  license:
    accept: true
    cloudPak: IBM Cloud Pak for Business Automation
  defaultStorageClassName: managed-nfs-storage
  taskmining: 
    install: true
    storage:
      database:
        external: 
          credential:
            secretname: 'c-db2ucluster-cp4ba-instancepassword'
            passwordkey: password
          user: 'db2inst1'
          database: CP4ADB
          host: c-db2ucluster-cp4ba-db2u.test-pm.svc
          port: 50000
          type: 'db2'
          schema: 'TMUSER'
          ssl: false