Both exist; they serve different purposes on CP4BA:
- Performance Data Warehouse (PDW) is still deployed with each Workflow Server (it backs the Process Portal / Workplace reports, KPIs, SLAs and the tracking groups). Tracking groups and auto-tracking work unchanged; the PDW database is one of the databases you provide. Direct SQL reporting on the PDW tables remains possible but is not where IBM invests.
- Business Automation Insights (BAI) is the Cloud Pak analytics layer: the workflow engine emits events (process started / ended, activity, tracking points, case events) to Kafka; BAI's Flink jobs process them into OpenSearch / Elasticsearch indexes and the Kibana / Business Performance Center dashboards; the raw events can also be landed in HDFS / object storage for a data lake. Enable it in the CR (sc_optional_components: bai and the workflow section's business event emitter settings) and the dashboards appear in the Cloud Pak.
spec:
sc_optional_components: baw_authoring,bai
bai_configuration:
business_performance_center: { replicas: 1 }
flink: { task_managers: 2 }
baw_configuration:
- name: bawins1
# business event emitter of the workflow server: enabled (field name per release) - emits process / activity / tracking-group events to KafkaRecommendation: keep tracking groups (they become BAI events too, with their tracked fields), model KPIs in Business Performance Center dashboards, and treat the PDW as the engine's operational store rather than a reporting database; when you need your own BI tool, consume the BAI indexes or the Kafka topics, not the PDW tables. Instance data for operational reporting stays reachable through the Process REST v2 / search APIs.
References