Prerequisites
Familiarity with your target runtime platform (Kubernetes, OpenShift, or Docker) and your organization’s deployment and DevOps practices. To build and run models locally, work through the environment setup first.Design the architecture
1
Choose an architecture style
Decide between the
Adaptive Process Architecture,
which co-locates the process components in a single Quarkus deployment, and an
event-driven architecture
built on Apache Kafka and CloudEvents. Base the choice on your scale,
coupling, and latency needs.
2
Model decisions and processes as services
Package decisions (DMN/DRL) and processes (BPMN) as independently deployable,
cloud-native services. The
integration patterns cover embedding
decisions in processes, event-driven messaging, and service orchestration.
3
Plan governance
Decide how models are reviewed, promoted between environments, and audited.
Decision Control centralizes deployment and
management of decision services, and
Decision Control Tower adds governance
workflows and an audit trail across environments.
4
Plan deployment
Select your deployment target (Kubernetes, OpenShift, or Docker) and your
multi-environment promotion strategy. See the
deployment guides and, for image and runtime choices,
JVM vs native.
5
Integrate with DevOps
Wire builds and deployments into your CI/CD and observability tooling, and
secure the architecture
before exposing it.