Separate Bee nodes for production and staging environments

[crtahlin]

Problem

Production (provenance-gateway.datafund.io) and staging (provenance-gateway.dev.datafund.io) both point to the same Bee node. This causes:

1. Stamp ownership overlap — stamps purchased or adopted by one environment's pool are visible to both. A stamp acquired on staging can appear in production's pool and vice versa.
2. Pool interference — both environments run sync_from_bee_node() against the same stamp set, potentially competing for the same stamps or double-counting reserves.
3. Unpredictable testing — staging stamp operations (purchases, uploads, pool drains) affect production stamp availability and BZZ balance.

Proposal

Each environment should connect to its own Bee node:

Environment	Bee node	Purpose
Production	Dedicated node (current)	Live users, real data
Staging	Separate node	Testing, validation before production

Implementation

1. Deploy a second Bee node for staging (or repurpose an existing test node)
2. Set SWARM_BEE_API_URL to the staging Bee node URL in the GitHub staging environment variables
3. Fund the staging node's wallet with enough xBZZ/xDAI for stamp purchases during testing
4. Verify staging pool operates independently after the switch

Considerations

• The staging Bee node doesn't need the same capacity as production — a lighter setup is fine
• Staging wallet needs its own xBZZ funding (separate from production)
• After the split, staging pool state should be reset (delete data/pool_state.json on staging) to avoid referencing stamps from the shared node