What network access is required from the operator/plugin namespace to the cluster namespaces

The docs recommend installing the plugin in the same namespace as the cnpg operator (https://cloudnative-pg.io/plugin-barman-cloud/docs/installation/). I'll call it cnpg-system.

It seems common to run cnpg clusters in other namespaces. E.g., there are some nice networking docs here: https://cloudnative-pg.io/documentation/1.26/networking/#cross-namespace-network-policy-for-the-operator

But I don't see any docs for the networking aspects of the plugin. Does the plugin communicate with the cluster pods? With the Backup resource? I imagine it would need to, but maybe not? In any case, I appreciate any tips.

Start with the plugin installation documentation and the referenced CNPG networking documentation. Determine the plugin's required network access between cnpg-system and cluster namespaces, including whether it communicates with cluster pods or Backup resources; done means the documentation clearly states the required connectivity and configuration.

Written by the indexing model from the issue text.