Deploying and executing decision services
Multiple deployment options are available to help you execute decisions at scale where you need them:
Building and deploying decision services
- Building and deploying in a CI/CD stack
-
Decision Intelligence can connect to existing tools, allowing organizations with established DevOps practices to automate builds, run tests, and deploy decision services using their own CI/CD stack. Decision services that are built in the CI/CD pipeline can then be deployed to the decision runtime using REST API methods.
If your team does not have its own processes and tools in place already, Decision Intelligence provides a sample CI/CD stack that can be easily installed and configured, primarily for evaluation and demonstration purposes.
- Building and deploying from Decision Designer
-
For simpler workflows, Decision Intelligence also allows you to build and deploy decision services directly from Decision Designer.
Both the CI/CD-driven deployment and Decision Designer-initiated deployment use secure mechanisms to ensure decision services can be promoted safely into production environments.
Executing decision services
The decision runtime is a Kubernetes‑based service for executing decision services and managing them in organized deployment spaces. Each space holds decision service archives and metadata, enabling environment‑specific configurations.
The runtime provides a set of REST APIs for execution, resource management, diagnostics, and traces. Each instance runs as a WebSphere Liberty pod, supports multiple decisions, caches artifacts for performance, and applies metadata updates automatically.
It scales horizontally through independent replicas and is accessed through the IBM Cloud Pak Platform UI (Zen) with SSO and API key support. The runtime can also emit business and technical events to Apache Kafka.