Pricing
Simple, predictable monthly pricing.
Scale your compute capacity as your data grows.
Bronze
A namespace on a shared test cell — one region, shared nodes, Kafka included and quota'd. Evaluation, dev and CI, not production.
Minimum 2 vNodes — $158.00/month to start. Up to 8 vNodes; above that, Silver is the right tier.
Includes:
- A namespace on a shared, platform-owned cell — shared nodes, VPC and brokers
- 2 vCPU + 4 GB RAM per vNode · 2 to 8 vNodes per environment
- 20 GB egress + 20 GB state storage per vNode · ingress free
- Kafka included and shared: 10 topics, 3 partitions each, 3 days / 5 GB retention, 2 MB/s in and 4 MB/s out
- Flink, Kafka Connect, dashboards and logs (kept 7 days) included
- Not for production: no data-plane SLA, and the shared Kafka's data is transient
- Bring your own S3 bucket for Flink state and logs, on any tier
- Spend cap $0 by default — a hard stop at the bundle, never an invoice you did not expect
Silver
A dedicated Kubernetes cluster in one region — Flink, Kafka, monitoring and state storage included. The production default.
Minimum 4 vNodes — $516.00/month to start. Scale one vNode at a time.
Includes:
- A dedicated Kubernetes cluster — never shared
- 2 vCPU + 4 GB RAM per vNode · from 4 vNodes (the SYSTEM pool alone takes 2)
- 50 GB egress + 50 GB state storage per vNode · ingress free
- Kafka optional and dedicated: a 3-broker pool, one per availability zone
- 99.5% platform availability · rebuild runbook, RTO ≤ 60 min, RPO = last checkpoint
- Dedicated monitoring stack; log retention 1–730 days in your own bucket
- Scale one vNode at a time; usage beyond the bundle is metered under a spend cap you set
- Zero lock-in — vanilla OSS Flink, your YAML, any source or sink
Gold
A dedicated cluster in each of two regions with one stretched Kafka — zero RPO for acknowledged messages, and a region failure your customers never notice.
Minimum 4 vNodes per region — $1,592.00/month to start. Scale one vNode at a time.
Includes:
- A dedicated cluster in EACH of two regions, plus a witness voter we host
- Priced per vNode per REGION — from 4 vNodes in each region
- 50 GB egress + 50 GB state storage + 100 GB cross-region replication per billed vNode
- One stretched Kafka cluster across the pair — zero RPO for acknowledged messages
- 99.9% availability · region-loss failover ≤ 15 minutes, with published drill reports
- Per-job region placement: leave every job on the primary, or spread them across the pair
- SLA credits, and the drill history to check them against
Questions & Answers
Have anymore questions?
Chat with someone on our sales team who can help you find the right plan for your needs.
Enterprise-ready • Isolated Infrastructure • 24/7 Priority Support