Network
DDoS L3/L4 protection
Always on, scrubbing up to ~600 Gbps at the edge. Per-VIP rate limits tunable per org.
Billed monthly, in EUR, against a signed quota and an overage cap of your choosing. We don't have starter plans, growth plans, or pro plans. We have one price sheet, one contract template, and three sensible cap values. The orchestrator reads the contract before it bills anything.
| Resource | Description | Price |
|---|---|---|
| vCPU | Per virtual CPU committed against quota, on KVM general-purpose hosts (1:4 vCPU:GiB sizing family). | €18.00per vCPU · month |
| RAM | Per GiB of committed memory across the org's quota, scoped to the same hosts as the vCPU. | €4.00per GiB · month |
| Block storage | Ceph RBD, 3× replication across zones a/b/c. Boot disks and persistent volumes both billed here. | €0.12per GiB · month |
| Object storage | Ceph RGW, S3-compatible, 3× replication. Inactive-tier (IA) shipping Q3 — half this price. | €0.012per GiB · month |
| Egress | Public internet egress. Included up to 5× the vCPU quota in GiB / month. Beyond that, billed per GiB. | €0.008per GiB · over inclusion |
| DDoS L3/L4 | Always-on. Up to ~600 Gbps scrubbing at the edge. Not a separately billed add-on. | included— |
| Per-org Prometheus | Scrape configs, dashboards, alertmanager templates. Retention & routing are yours. | included— |
| Postmortems | Public within 10 business days of any incident touching paying orgs. | included— |
A single transactional service, low burst, one engineer on call. Typical for first-year tickets & bookings products.
The shape most paying orgs land on: predictable monthly load, two evening spikes, one on-call rotation, one ADR file.
100+ vCPU committed, dedicated hosts on opt-in, custom cap. Quarterly business review with platform-eu.
Always on, scrubbing up to ~600 Gbps at the edge. Per-VIP rate limits tunable per org.
Scrape configs, dashboards, alertmanager templates. Retention is yours; we don't bill for it.
Copy-on-write block snapshots, per-pool retention you set in the contract. Off-cluster export is on the roadmap.
Anything that touches paying orgs: postmortem within 10 business days, public on /engineering/, with the dashboard we kept open.
All data stays in nl-ams-1. DPA & GDPR posted, no transfers to third countries.
Slack Connect or signed email with the engineer who would run your envelope. Not a tiered support contract.
Every scale event signed (ed25519) and appended to your org ledger. Pulled by the API; mirrored monthly to RGW IA tier.
Go, Python, a CLI, a Terraform provider. Same surface as the console; nothing held back.
HTTP 429. Your application returns whatever it would return during any over-capacity event. We email the org owner within 30 seconds, append a refusal line to your ledger, and the platform stops growing the pool. We do not phone you and we do not silently invoice the overage. Raising the cap is a co-signed addendum, applied at the start of the next billing month.rbd export, s3cmd sync). We can hold a final snapshot for 30 additional days at the standard block-storage rate.