[capistrant]

Description

• When using druid-kubernetes-extensions for node discovery, the K8s watch loop replays ADDED events for existing pods on every reconnect. This caused BaseNodeRoleWatcher.addNode() to log at ERROR level for each
  duplicate, polluting logs with non-actionable noise.
• Downgraded the "discovered but existed already" log from ERROR to DEBUG, since duplicate adds are expected behavior during watch reconnects.
• Moved the INFO "Node detected" log into addNode() so it only fires for genuinely new nodes, not duplicates.
• Added a test covering the reconnect scenario (resetNodes followed by duplicate childAdded calls).

Release note

N/A

---

Key changed/added classes in this PR

• BaseNodeRoleWatcher

---

This PR has:

• been self-reviewed.
• added documentation for new or modified features or behaviors.
• a release note entry in the PR description.
• added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
• added or updated version, license, or notice information in licenses.yaml
• added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage is met.
• added integration tests.
• been tested in a test Druid cluster.