Operational Data Readiness
Use this guide before a Data Center Operations dashboard, diagnosis workflow, or dispatch review becomes part of daily operations. The page helps teams separate reviewed operating data from values that still need mapping or source validation.
Readiness Flow
Prerequisites
| Requirement | Why it matters |
|---|---|
| Source owner | Each source needs someone who can explain fields, units, and sync behavior. |
| DFS or connector path | Mappings, sync history, and rejected records need a governed review path. |
| Review cadence | Operations teams need to know how often data readiness is checked. |
Inputs
| Source | Readiness checks |
|---|---|
| Asset inventory | Stable IDs, hierarchy, criticality, owner, and location. |
| BMS | Point mapping, unit, scale, timestamp, source freshness, and mapping approval. |
| Alerts | Severity, status, affected equipment, source timestamps, and repeat pattern. |
| Work orders | Owner, status, priority, action notes, feedback, and closure evidence. |
| Meters | Utility type, scope, unit, time window, and coverage. |
| Predictive signals | Health snapshots, prediction history, model version, and signal freshness. |
Procedure
- Inventory the source systems used by the first operating boundary.
- Confirm the required fields for dashboard, asset, diagnosis, and dispatch views.
- Review sync history and rejected records.
- Check timestamps, units, scale, and timezone handling.
- Mark each source as ready, ready with limitation, or blocked.
- Record limitations in the operating handover.
Expected Output
The team should have a readiness table showing which data can support daily review, which data can support evidence only, and which data must be corrected before use.
Validation Checklist
- Source freshness is visible for every active feed.
- Units and scale are documented for BMS and meter values.
- Asset IDs are consistent across alerts and work orders.
- Predictive outputs link to source signal freshness.
- Limitations are visible in the review record.
Failure Handling
| Symptom | Response |
|---|---|
| Values are missing for a time window | Check connector sync, source availability, and rejected records. |
| Units differ across sources | Normalize units in mapping and document the conversion. |
| Work-order feedback is incomplete | Keep closed-loop reporting in review mode until feedback fields are populated. |
| Predictive signal is stale | Do not use the prediction for dispatch until the data owner confirms freshness. |