ClusterCheck: The `empty_default_hostname` option has no effect in the Postgres check.

[dimke931]

I am using the latest agent version 7.73.1, and in cluster checks I see the following warning:

2025-12-17 18:48:54 UTC | CORE | WARN | (pkg/collector/python/datadog_agent.go:145 in LogMessage) | unknown | (postgres.py:136) | The empty_default_hostname option has no effect in the Postgres check. Use the exclude_hostname option instead.

I cannot find where this parameter is configured explicitly.
After investigation, I found that it seems to be injected by the Cluster Checks code here

When I run agent configcheck postgres -v, I see empty_default_hostname: true in the generated configuration:

agent configcheck postgres -v

heck_name: postgres
init_config: {}
instances:
- custom_queries:
  - columns:
    - name: blocked_query
      type: gauge
    metric_prefix: postgresql
    query: my_query
  - columns:
    - name: corrupted_index
      type: gauge
    - name: database_name
      type: tag
    - name: index_name
      type: tag
    metric_prefix: postgresql
    query: my_query
  dbname: DBNAME
  dbstrict: "true"
  empty_default_hostname: true
  host: DBHOST
  password: '********'
  port: "5432"
  query_timeout: "1000"
  relations:
  - relation_regex: .*
  tags:
  - tag:tag
  username: datadog
logs_config: null

My understanding is that this parameter is automatically added by the Cluster Agent for cluster checks, but the Postgres integration does not support it and logs a warning instead.

Could you please confirm whether this is expected behavior or a bug?