Enabling tracing for BPM event emitter and Machine Learning Server

To help you troubleshoot any issue that might occur with the BPM event emitter and Machine Learning Server, you can enable detailed tracing by adding a package to the trace definition in the WebSphere® Application Server administration console.

About this task

You can trace the following activities for the BPM event emitter and Machine Learning Server.

BPM event emitter:
- Whether the native JSON DEF events are correctly transformed to IBM Business Automation Workflow raw events.
- Whether the transformed Business Automation Workflow events have successfully been sent to Kafka.
New in 20.0.3 Machine Learning Server:
- Whether the requests to Machine Learning Server are successful.
- Whether the response from Machine Learning Server is correct.

Procedure

Follow the correct set of steps based on whether Machine Learning Server is on premise or in containers.

If you are troubleshooting BPM event emitter and Machine Learning Server for IBM Business Automation Workflow traditional (on premise) installation, see the following steps:
1. Open the WebSphere Application Server administration console.
2. Navigate to Troubleshooting > Logs and trace > application_cluster_name > Change log details level > Runtime.
  
  The application_cluster_name placeholder is a generic name that is provided at Business Automation Workflow installation time.
3. Add the appropriate package to the trace definition. The trace definition setting is dynamic, so you do not need to restart the cluster members after adding the package.
  - To troubleshoot issues with the BPM emitter, add the following package to the trace definition:
    : com.ibm.bpm.mon.oi.*=all
  - New in 19.0.1 To troubleshoot issues with playing back events from Performance Data Warehouse, add the following package to the trace definition:
    : com.ibm.bpm.bui.emitter.*=all
  - New in 20.0.3 To troubleshoot issues with the Machine Learning Server, add the following package to the trace definition:
    : com.ibm.bpm.bui.ml.*=all
4. For each application cluster member, the trace is written to the BAW_Install_Dir/profiles/node_profile/logs/cluster_member/trace.log file.
If you are troubleshooting BPM event emitter and Machine Learning Server for Cloud Pak for Automation, see the following steps:
1. Add the trace definition by appending to baw_configuration[x].logs.trace_specification in the custom resource definition:
```
baw_configuration:
  logs:
    trace_specification:
      ": trace_string_1: trace_string_2"
```
2. Add the appropriate package to the trace definition. The trace definition setting is dynamic, so you do not need to restart the cluster members after adding the package.
  - To troubleshoot issues with BPM emitter, add the following package to the trace definition:
```
: com.ibm.bpm.mon.oi.*=all
```
  - To troubleshoot issues with playing back events from Performance Data Warehouse, add the following package to the trace definition:
```
: com.ibm.bpm.bui.emitter.*=all
```
  - New in 20.0.3 To troubleshoot issues with the Machine Learning Server, add the following package to the trace definition:
    : com.ibm.bpm.bui.ml.*=all
3. Apply the custom resource .yaml file with the new trace definition:
```
oc apply -f cr.yaml
```
4. For each Business Automation Workflow instance, the logs can be retrieved by accessing the Business Automation Workflow pod logs.
  1. Find the Business Automation Workflow instance pod name:
```
oc get pods | grep baw
```
  2. View the pod log for each Business Automation Workflow instance:
```
oc logs -f baw_instance_pod_name
```