Infrastructure cost
The marginal event
costs zero.
You pay for hardware you already control — nothing else. No per-event metering, no per-site fees, no per-seat math. At roughly 60 million events a month, the whole bill is the box, the disk and a few euros of object storage.
The bill
One line item nobody can inflate.
The mockup below is your invoice, not ours — every line is infrastructure you own, sized together with your DevOps during onboarding.
| VPS — 2 vCPU / 4 GB | €6–15 |
| Object storage — backups + snapshots | €2–5 |
| ClickHouse disk headroom (flat by TTL) | €0–20 |
| Observer license | €0 |
| Per-event charges | €0 |
Estimates from typical European VPS and object-storage pricing; your numbers depend on the provider you already use. The Observer license is negotiated once, after NDA — it never scales with traffic, sites or seats.
Why it stays flat
Metering is structural — not your problem here.
SaaS metering is structural
SaaS vendors meter events because their compute scales with your traffic — the bill grows because their costs do. It is not greed; it is their architecture. Yours is different.
Your headroom is the buffer
The 60-millionth event in a month costs the same as the first: whatever headroom your server had. A viral week costs disk headroom, not a budget exception.
Flat as you grow
Retention is TTL-enforced and the heavy lifting happens in rollups, so growth shows up as CPU you can add — not as a line item that doubles when your launch succeeds.
Against the meter
Same events, different curve.
On a metered SaaS, cost tracks traffic linearly forever — the pricing page decides what a successful year costs you. Here, the curve flattens at the price of a small VPS, because the machines are yours. The license is agreed once; the events do not vote on it.
Growth-proof
by architecture.
Bring a €6-class VPS and the stack runs; bring 60 million events and it still runs. The marginal event is free because your hardware, not a meter, decides the bill.