LFA · capacity model

Fidelity & Data-Point Calculator

Model how many telemetry data points an LFA run stores — full 15-minute fidelity versus a policy that keeps older data in coarser time slots, while coarser rollups stay alongside the finer recent data.

Full fidelity Aggregated
01

Network & metrics

Size of a single network
Run scale
Metrics stored full fidelity
Metrics stored every aggregated window
Per 15-min slot · one network: · all networks: Per aggregated slot · all networks:
02

Aggregation policy

Each rule says: data older than X months is also kept in a coarser time slot. A coarser slot is retained from now back to the age where data drops below it — so recent periods hold the fine data and every coarser rollup at once. Rules apply by age, youngest first; every aggregated window uses the same metric counts set above.

BASE
15-min fidelity
full-fidelity metrics · recent window
03

Result

over 60 months
Reduction vs full fidelity
Full fidelity
With aggregation
Full: Agg:
04

How it's calculated

your numbers, step by step
05

Resolution layers

nested · each kept back to here

Each resolution is stored from now back to the age where data becomes coarser than it, so the finer recent periods also hold every coarser rollup. Bar length = how far back that resolution is kept across the retention window.

now
06

Per-resolution breakdown

ResolutionNode m.Edge m.Kept to SlotsData points% agg
Total 100%
Model · for each resolution: slots = (months kept × 43,200) ÷ slot-minutes, rounded; points = slots × networks × (nodes·node-metrics + edges·edge-metrics). Totalled across all resolutions.
Aggregated slots · the node/edge metric counts shown for coarser resolutions are the effective stored counts: aggregated set + a full-fidelity min set + a full-fidelity max set.
Coverage · a resolution is kept from now back until data drops below it (the next coarser rule's threshold), so coarse rollups overlap the finer recent periods.
Time basis · 1 month = 30 days, 1 year = 365 days.