Pay for what you use: Cube Cloud pricing gets back to its roots

Cover of the 'Pay for what you use: Cube Cloud pricing gets back to its roots' blog post

We launched Cube Cloud back in 2021 and have been busy ever since launching new features, growing Cube Cloud usage, and listening to your feedback. Today, we are happy to make a major update to Cube Cloud pricing.

This update—measuring usage with Cube Compute Units—makes pricing more predictable and reflects the diversity of enterprise customers using Cube Cloud with custom-tailored pricing plans. Let’s walk through these changes.

Transparent and predictable pricing

We have always wanted Cube Cloud pricing to be transparent and predictable for Cube Cloud users, so that it reflects the actual usage of the service and can be forecast in advance. That is why we think that consumption-based pricing is the best option for Cube Cloud.

Initially, we chose data pass through—defined as the amount of data transferred between Cube Cloud and API clients or database—as the measure of resource consumption. It definitely helped users quantitatively understand how much data flows from a data warehouse to Cube and from Cube to data consumers.

However, this model turned out to not be transparent and predictable enough. Often, the most pragmatic way to understand the expected data pass through of a deployment was to actually make it live and then measure. Also, changes to data model, caching configuration, queries, etc. would sometimes result in step changes and make data pass through hard to predict. For these reasons, we’ve decided to introduce a better alternative.

Cube Compute Units (CCUs)

Now, our users’ resource consumption is measured in Cube Compute Units, which precisely and transparently reflect the compute required to run certain Cube components, i.e., API instances and development instances, refresh workers, Cube Cloud routers and workers. (Note that Redis instances are not present in this list since we have migrated from Redis to Cube Store.)

Cube Compute Units

Let’s explain. Assume that you have a minimal yet very potent Production Deployment that includes two Cube API Instances (enough for serving up to 20 requests per second), a Cube Refresh Worker, and two Cube Store Routers. You also choose to include two Cube Store Workers because you want to use pre-aggregations and provide your users with blazing-fast experience:

Architecture

We can go through a few real-world scenarios for that deployment:

  1. Running this deployment would consume 4 to 6 CCUs per hour or approximately 2900 to 4400 CCUs per month (keep in mind that when pre-aggregations are not being refreshed or used to serve queries, Cube Store Workers are scaled down).
  2. Additionally, if every workday at peak hour your API autoscales from 2 to 10 instances, that would add roughly 200 CCUs per month.
  3. Also, if two engineers on your team hack on the data model for a couple of hours every workday while running their Development Instances, that would add about 100 CCUs per month.

For simplicity, we used whole hours in the calculation. However, Cube Cloud will actually scale instances up and then back down within minutes, so you’re never overcharged for the unused resources.

Updated Billing Experience

We believe that this flavor of consumption-based billing, popularized by many industry leaders such as Snowflake (credits) and Databricks (DBUs), is already well-known to many Cube Cloud users and the overall data community.

To further help Cube Cloud users understand and forecast their usage and related costs, we also updated the Billing page in Cube Cloud. Now, you’re able to break down usage by deployments and services, review it on a daily and monthly basis, and monitor spend in CCUs and USD:

Billing interface

Needless to say, this page is powered by Cube and provides near real-time experience.

New plans for production deployments

Over the last couple of years, we’ve accumulated vast knowledge of Cube Cloud users. Our familiarity with our user base provided us with the insights to launch many new features and update pricing plans so they’re tailored to their everyday needs.

You can see the entirety of new plans on the updated pricing page.

Try a Development Instance of Cube Cloud, free of charge

One hill we’ll die on? Using Cube Cloud is the best way to get started with Cube.

Fully managed, secure, and packed with features that provide a smooth developer experience, Cube Cloud should be available to every single engineer getting started with Cube. So, in that vein,, we continue to provide an option to try Cube Cloud for free for your non-production or proof-of-concept deployment.

This option empowers you to evaluate Cube indefinitely and upgrade to any of the following plans when you’re ready to go into production.

Choose a plan for production deployments

We’re introducing four new plans: Starter, Premium, Enterprise, and Enterprise Premier.

Every plan has a different price per CCU—and for a good reason. We’ve built these plans so they reflect the needs of different Cube Cloud users: from seed-stage startups to Fortune 500 companies with mission-critical data applications.

A few highlights regarding these new plans:

  • Starter plan will let you run deployments on AWS and GCP, collaborate as a team using development mode and data model IDE, and have full observability with real-time application performance monitoring (APM) and alerts.
  • Premium plan guarantees 99.950% uptime, lets you run deployments in a dedicated virtual private cloud (VPC), supports Azure in addition to AWS and GCP, and provides your team with single sign-on (SSO) via Okta.
  • Enterprise plan guarantees 99.990% uptime, provides VPC peering, logs and metrics export, SSO via SAML/LDAP as well as role-based access control (RBAC). You will also get a dedicated customer support manager (CSM) who would be in direct contact with you at every step of your Cube Cloud experience.
  • Enterprise Premier plan guarantees 99.995% uptime, provides unlimited scalability with multi-cluster deployments as well as connectivity to Elastic and Kafka (via ksqlDB) for real-time streaming. You will also get the utmost level of support for your mission-critical data applications.

Indeed, Cube Cloud is packed with features for organizations of all sizes, so one could get overwhelmed with the available options—but don’t fret. We’re always happy to help out, so just get in touch with us.

What’s next

Our updated pricing with new plans, Cube Compute Units, and great billing experience are already available to all new Cube Cloud users. Check our updated pricing page and start using Cube Cloud today. Also, don't hesitate to talk to us so we can help you get started.

Note that nothing is changing for existing Cube Cloud users on Standard and Enterprise plans as of right now. If you’re in this amazing category, thank you for your support. If you see something in the new plans that gets you excited, please reach out and we will get you set up. Or—just carry on using Cube Cloud on your current plan and we’ll get in touch sometime soon to assist you with upgrading to new plans.

If you have any questions or feedback regarding our updated pricing, please reach out; we’d be happy to walk you through how it works. Or share your thoughts in the #cube-cloud channel in our Slack community—because even with almost 8,000 data practitioners, it’s still never too late to join 🙂

Share this article