Issue: Implementation of a Jira DC cluster requires a load balancer to store a persistence record of a users session so requests with the users session ID are routed to the specific Jira DC node where that session is present. In turn, an administrator is obliged to maintain this configuration on the load balancer as part of their Jira DC implementation. This is undesirable as it has both an implementation and operational overhead in addition to users being logged out if a node fails or is taken offline for maintenance.

Goal: Session ID's are shared between Jira DC nodes eliminating the need for persistence records on a load balancer and in turn allow any http request to be served from any Jira DC node.

Benefits:

Reduction of Jira DC implementation complexity.

Jira DC nodes can fail or be disabled without users loosing their sessions.

Enhanced HA.