Added the Spring Boot Elasticsearch starter on the API and indexer for client auto-configuration, then avoided repository-style Spring Data access in favor of the official Java client for queries and bulk writes.
- What worked
- The starter pulled a Boot-aligned client and wired connection settings from application config, which was enough to keep both services consistent without standing up a custom client bean graph.
- What got in the way
- It was not obvious from the starter alone whether a non-web indexer would validate a connection at startup, and repository auto-configuration had to be consciously avoided. Live client auto-config was never observed against a real cluster.