Add details to why the minio tenant headless service must be used over the normal clusterIP service

["The console service is for managing the tenant, while the hl service exposes the S3 compatible API. "](https://cloudnative-pg.io/plugin-barman-cloud/docs/object_stores/#minio-object-store)

the `-hl` service is a headless service. There is also a `minio` "normal" (clusterIP) service created. Why can't we used the "normal" service? Does the plugin have to connect to all of the pods behind the service? I have been testing with a single pod and it works with the normal service (not the headless version) but that may just be because there is a single pod. Can you update the docs to say why it can't use the normal service, if indeed that is the case? Thanks!

EDIT: Furthermore, how does https://cloudnative-pg.io/plugin-barman-cloud/docs/retention/ ("The Barman Cloud Plugin supports automated cleanup of obsolete backups via retention policies, configured in the .spec.retentionPolicy field of the ObjectStore resource.") square with https://cloudnative-pg.io/plugin-barman-cloud/docs/object_stores/#s3-lifecycle-policy ("Barman Cloud uploads backup files to S3 but does not modify or delete them afterward. ") ?

Start with the linked MinIO object stores and retention documentation pages. Verify why the tenant headless service is required compared with the ClusterIP service, and reconcile the retention-policy wording with the S3 lifecycle-policy statement. Done means the documentation clearly explains the supported service and presents consistent retention behavior.

Written by the indexing model from the issue text.