Multi-Pod Cluster Deployment

Multi-pod cluster deployment is more complex than single pod deployment. It is also more resilient to hardware failures and is suitable for horizontal scaling.

TIBCO PSI attains high availability in this deployment due to multiple instances of TIBCO PSI servers in different pods. If a pod crashes and stops working, the remaining pods in TIBCO PSI service are still available to service the incoming requests. The remaining pods continue to function until a replacement pod is started by the Kubernetes master. The data remains safe as the database runs on a different pod under a different exposed service.

Note: To make a database (PostgreSQL or Oracle) or server (TIBCO FTL or Enterprise Message Service™) highly available, you must deploy them in high availability mode. For more information on the high availability deployment of PostgreSQL, Oracle, TIBCO FTL, or Enterprise Message Service™, refer to the respective documentation.

The TIBCO PSI GUI is accessed using the IP address of TIBCO PSI service. TIBCO PSI service is a Kubernetes object. TIBCO PSI service routes the request to an available PSI Pod, for example, http://<ip address of PSI Service>:8080/psi-server.