ScyllaDB publishes its pricing structure on scylladb.com/pricing. Unlike per-request models, ScyllaDB Cloud bills based on resources you provision: instance type (vCPU and RAM), NVMe SSD storage, service tier, deployment model, and variable cloud costs for network and backups.

For a monthly estimate, use the Pricing Calculator linked from the pricing page as "Estimate Costs."

How ScyllaDB Cloud billing is calculated

Per ScyllaDB's pricing FAQ, billing is based on:

ScyllaDB states that pricing is "transparent and based on the resources provisioned," not a black-box serverless model.

Service tiers: Standard, Professional, Premium

ScyllaDB Cloud offers three tiers. Feature lists below come from scylladb.com/pricing.

Standard

• 2 Hour 8x5 P1 support response
• Fully managed on AWS and GCP
• Vector Search, X Cloud auto-scaling storage and compute
• SOC2 and ISO certifications
• Pricing page hero card lists 99.99% uptime SLA; FAQ lists 99.9% uptime for Standard tier production workloads

Professional

All Standard features, plus:

• 1 Hour 24x7 P1 support response
• Bring Your Own Account (BYOA)
• Multi-region active-active HA/DR
• Advanced metrics and logging
• Dedicated Slack channel
• Customized maintenance windows

Premium

All Professional features, plus:

• 15 minutes 24x7 P1 support response
• Bring Your Own Encryption Key (BYOK)
• Advanced networking (Private Link)
• Advanced network compression
• Flexible region switching
• Dedicated Technical Account Manager
• Custom licensing models
• FAQ lists 99.99% uptime SLA for Premium large-scale enterprise environments

ScyllaDB notes you cannot unbundle cloud product and support tiers: upgrading one upgrades both.

Billing models on ScyllaDB Cloud

ScyllaDB's pricing FAQ describes three billing approaches:

On-Demand. Hourly rate for resources with no long-term commitment. Suited to unpredictable workloads or short-term projects.

Subscription (Fixed Contract). Commitment for 1-year or 3-year terms. ScyllaDB states savings of "up to 70% or more" compared to On-Demand rates for steady-state production workloads.

Flex Credits. Prepaid pool of credits at a discounted rate, applied to dynamic usage such as temporary scaling during traffic spikes or adding nodes. ScyllaDB describes this as a middle ground between On-Demand and fixed contracts.

Enterprise and self-managed ScyllaDB deployments use "Get a Quote" pricing on the same page. ScyllaDB Enterprise runs on public cloud, private cloud, or on-premise with enterprise service and support options.

What ScyllaDB does not charge for

ScyllaDB's FAQ states it does not charge per individual request (read/write units) like DynamoDB. You get the throughput capacity of the provisioned hardware. ScyllaDB claims this "often results in a 50% or greater cost reduction compared to DynamoDB for high-throughput workloads" on its pricing page.

Use the official calculator with your workload parameters rather than extrapolating from marketing comparisons.

Free trials

ScyllaDB Cloud offers two free entry paths, per its pricing FAQ:

• Developer Free Trial: 30-day trial on smaller instances (e.g., AWS t3.micro) for exploring CQL and prototypes
• Production Evaluation: 48-hour trial of production-grade instances for benchmarking and latency testing

BYOA deployment model

Bring Your Own Account (BYOA) runs the DBaaS deployment in your AWS or GCP account. You pay the cloud provider directly for infrastructure; ScyllaDB charges a management fee while its SRE team manages the database. BYOA is available on Professional and Premium tiers per the pricing page feature comparison.

Multi-AZ, multi-region, and variable costs

Multi-AZ. ScyllaDB states clusters are typically distributed across three AZs. You pay for instances and storage in each zone, plus standard cloud provider data transfer fees for inter-zone replication.

Multi-Region. You pay for resources in each region, plus standard cloud provider data transfer fees for inter-region replication. ScyllaDB notes rack-aware drivers can reduce cross-zone and cross-region costs.

Data transfer. Standard cloud provider rates apply for internet egress or cross-region traffic.

Backups. One daily backup is included in the managed service. Additional backup frequency or long-term snapshot retention may incur extra storage fees in S3 or GCS.

ScyllaDB X Cloud

ScyllaDB describes X Cloud on its pricing page as an "elastic database designed to support variable/unpredictable workloads with consistent low latency as well as low costs." X Cloud auto-scaling storage and compute appears on the Standard tier feature list.

The analytics cost consideration

ScyllaDB's pricing page describes a wide-column NoSQL database optimized for partition-key access patterns, high throughput, and low latency on keyed reads and writes. It does not position ScyllaDB as an OLAP engine.

When teams run cross-partition aggregations, dashboards, or ad-hoc analytical scans on ScyllaDB, they pay for provisioned compute and storage on workloads the data model was not designed for. Denormalizing tables for each query pattern increases storage under the same billing model.

For DynamoDB real-time analytics, the same architectural split applies to ScyllaDB: keep the operational database for keyed workloads, stream changes to an analytical engine for aggregations.

ScyllaDB (OLTP) → CDC → Kafka → ClickHouse®/Tinybird (OLAP) → SQL APIs

ScyllaDB documents Change Data Capture (CDC) on its pricing page for tracking changes to any table in any keyspace. For real-time data ingestion into an analytics layer, CDC via Kafka is the standard pattern.

Pairing ScyllaDB with an analytics layer

ScyllaDB bills for provisioned vCPU, RAM, and NVMe storage. Analytical workloads that scan across partitions consume that capacity inefficiently compared to a columnar OLAP engine.

For ClickHouse Cassandra integration, the hybrid pattern keeps ScyllaDB costs aligned with operational access patterns while ClickHouse handles aggregation queries.

For ScyllaDB alternatives, teams often search when analytics performance on ScyllaDB becomes the bottleneck, not when keyed OLTP performance fails.

For best database for real-time analytics, the operational and analytical layers typically use different engines priced for different access patterns.

For event sourcing with Kafka, CDC from ScyllaDB into Kafka feeds downstream analytics without dual writes to two operational stores.

Tinybird for ScyllaDB analytics

Tinybird is managed ClickHouse with streaming ingestion for teams running ScyllaDB as their operational database:

1. Enable ScyllaDB CDC to Kafka (ScyllaDB documents CDC and a shard-aware Kafka connector on its pricing page).
2. Connect Tinybird's Kafka datasource to the change stream.
3. Define SQL Pipes for dashboard and API queries.
4. Publish Pipes as HTTP endpoints with JWT tenant isolation.

ScyllaDB continues handling keyed reads and writes. Tinybird handles analytical queries that would otherwise require denormalized ScyllaDB tables or full scans.

Resend, processing 100TB per month on Tinybird, measured 62ms p90 query latency in production without caching. Tinybird is SOC 2 Type II certified.

What to verify before budgeting

ScyllaDB does not publish fixed per-vCPU dollar rates on its pricing page. To estimate cost:

1. Use the Pricing Calculator linked from scylladb.com/pricing.
2. Choose tier (Standard, Professional, Premium) based on support, BYOA, and compliance needs.
3. Choose billing model (On-Demand, Subscription, Flex Credits) based on workload predictability.
4. Account for multi-AZ replication (three AZs), multi-region data transfer, and backup retention beyond the included daily backup.

For operational keyed workloads, ScyllaDB's provisioned-resource model and lack of per-request charges are the core pricing advantages ScyllaDB itself highlights. For analytics, pair ScyllaDB with ClickHouse via Tinybird rather than expanding ScyllaDB storage and compute to cover OLAP access patterns.