Background Indexing vs Full Re-index for High-Availability Maintenance

When performing maintenance on a Jira Data Center instance, the primary goal is to rebuild the search index while minimizing the impact on end-user availability. There are two primary documented paths: Background Indexing and Full Re-indexing.

Background Indexing allows users to continue accessing issue data by building a new index in a staging area, whereas a Full Re-index requires maintenance mode, which blocks user access entirely. However, Background Indexing introduces specific resource constraints, notably requiring significantly more disk space and CPU overhead to manage concurrent data changes.

Given a scenario where system uptime is critical but hardware resources are strictly capped, the trade-off between total clock time and user availability becomes a primary design decision.

• Does the reduction in downtime provided by Background Indexing outweigh the risk of performance degradation for active users under high CPU load?
• At what specific ratio of available disk space to current index size does Background Indexing become an unstable choice compared to a scheduled Full Re-index?