Waku Store retention limits and pruning configuration for long-running nodes

Waku Store provides historical message retrieval from a local database on full nodes. The protocol defines the store query interface but does not mandate retention duration, pruning intervals, or resource caps. Retention, pruning and max peer settings are implementation choices, so behavior differs across clients and deployments.

The design goal is to operate a long-running full node that serves relay and store access to light clients without unbounded memory and disk growth. There is no protocol-level permission boundary for who may query the store or under what rate and resource constraints. Light clients depend on full nodes for relay and store access, leaving the permitted serving relationship unresolved.

What retention and pruning configuration options are documented for current Waku Store implementations? What permission boundary, if any, exists for light clients accessing store data from a full node? Are per-peer rate limits or resource caps defined for store queries?