XDL Data Flow Timing Flashcards
(20 cards)
Three freshness tiers in XDL?
Near-real-time • Intra-day • Daily batch.
Latency for near-real-time tier?
5–10 minutes after SA receives data.
Main tables in near-real-time tier?
*_intervals, *_registers, *_events, device_group, account.
How to eyeball freshness of a table?
Query MAX(ingest_timestamp) for that table.
Latency goal for customer-portal data?
≈1 hour to XDL.
Which tier is *_consumption?
Sub-daily batch (~4-hour latency).
RNI ‘Consumption Aggregator’ schedule?
Runs every 4 hours, then pushes to SA & XDL.
Latency for daily batch tables?
Available next calendar day before 04:00.
Give two daily-batch tables.
metrology_device, fl_logs (FieldLogic).
Why 10-minute target is ‘good enough’?
Meters cache locally; alarms within 1 polling cycle.
What does SA stream every 5-10 min?
Intervals, registers, events, CIS deltas.
Key control column to check freshness?
ingest_timestamp (or read_time_timestamp_local).
Who drives near-real-time push?
Sensus Analytics micro-batch bus.
Why not stream manufacturing data?
Rarely changes; daily batch saves bandwidth.
FieldLogic AMR read cadence?
Daily pull into XDL.
Portal_activity table freshness?
Hourly—good for same-day customer trends.
Which table for outage pings?
electric_events (near-real-time).
Which table shows firmware version?
metrology_device_history (daily).
Default nightly ETL start time?
≈02:00 local utility time.
One-liner to answer ‘How fresh is XDL’?
Intervals <10 min, consumption ≤4 h, metadata next-day.