Warehouse Dashboard Bug Fixes - Urgent Project

Our Warehouse OS Operations Console has grown quickly and a few cracks are starting to show. The most pressing pain-point is the dashboard: figures drift from reality, some widgets go blank after long shifts, and operators on the warehouse floor occasionally see numbers that don’t match what our API returns. Because the warehouse team makes real-time decisions from those screens, I want to stabilise dashboard data accuracy first, then move on to the remaining back-end and UI clean-up work already scoped.

Current stack & context
– Node/Express services expose REST + WebSocket feeds
– PostgreSQL with logical replication to a read replica for the console
– React + Redux front end wrapped in an Electron shell for the large-format terminals
– Redis and Bull queue for product file loading and issuance workflows

Key objectives you’ll tackle in this first phase
1. Trace and eliminate the drift between inventory counts in the DB and the dashboard tiles shown to Warehouse operators.
2. Harden the data pipeline so every dashboard refresh (manual or auto) reflects the last committed transaction with sub-second latency.
3. Add granular logging and validation so any future mismatch is captured with context (user id, replica lag, cache state).
4. Respect role considerations: Warehouse operators should never see partial or stale information that an Admin might tolerate for audit work.

If time allows in the same sprint we can also start closing related tickets around Product ID / ContentHash generation, device-bound issuance quirks, and SKU synchronisation, but the dashboard fix remains milestone #1.

What I need from you
• A quick walk-through of similar real-time dashboard or admin-panel issues you’ve solved (links or repo snippets are fine).
• Your preferred tech stack or any tooling you’d introduce (SQL profiler, Grafana, Kibana, Jest, Cypress, etc.).
• An outline timeline with checkpoints—especially how long you expect until we see the first verified “no-drift” dashboard refresh.
• Anything you need from me (schema diagrams, staging credentials, data samples).

Deliverable acceptance criteria
– 100 % match between API JSON payload and rendered dashboard values during a 48-hour stress test.
– Replica lag alerts fire when latency exceeds 500 ms.
– Warehouse operator role verified in staging with accurate, complete data across shift changes.

I’m happy to extend the collaboration into the broader backlog once the console proves rock-solid again.