Log Retention Size
Storage for logs from event rate and size over hot and cold retention windows, with compression, index overhead, replicas, daily growth and a cost per GB-month.
Raw GB per day and month, stored GB per day and in total for each tier, the retained total at steady state, and the monthly cost from the prices you enter — every factor yours.
Example: 500 events/s at 400 B is 17.3 GB a day raw; 7 hot days at 4:1 with 20 % index and 2 replicas hold 72.6 GB, 90 cold days at 8:1 hold 194.4 GB — 267 GB in all.
Volume × factors × days.
How the stored size is built from the rate, the tiers and the factors you supply.
Raw volume
Raw bytes per day = events per second × bytes per event × 86,400. That is the line or document size before anything a store does to it; the average event size and the rate are yours to measure from a day of real traffic.
Tiers
Stored per day = raw × (1 + index overhead) ÷ compression ratio × replicas — the index overhead for a searchable hot tier, the ratio the store actually achieves on your format, and every copy kept. Hot days use the hot factor, the cold or archive days after them use theirs, and the retained total is the sum at steady state, when a day expires as a new one lands. A cost per GB-month per tier turns that into a bill.
Growth
A daily growth percentage compounds across the window: the newest day carries (1 + g)^(days − 1) times today’s volume and the total is the sum, not days × today. Snapshots, merge temporaries and metadata are not included; compression, overhead and replica counts are not assumed — enter the values from your own cluster. Nothing leaves the browser; the same four anonymous usage counts as the rest of the site apply.
SOURCES
- raw per day = events/s × bytes × 86 400; stored = raw × (1 + index overhead) ÷ compression × replicas × days; growth compounds per day across the window; steady state, no snapshots
Last reviewed 21 September 2026. How results are checked: How we verify.