About the company

Industry: mobile developer tooling
Company size: 201-500

Bitrise is a company that focuses on CI/CD problems in mobile applications. A true startup, Bitrise was founded in 2014 and is the first (and as of yet, the only) Hungarian company to have made it into the Y Combinator startup accelerator program. The company is a heavy GCP user, with more than 200 employees around the world working in a fully remote setup.

Bitrise provides a Mobile DevOps and Continuous Integration & Delivery solution for faster, better and more efficient app development.

Cloud transparency for cost control

Bitrise was Rabbit’s first customer. Using Rabbit, Bitrise can get actionable recommendations to identify and optimize unnecessary or unused (and therefore, wastefully maintained) resources.

Such suggestions may include:

• Stopping virtual machines,
• Changing machine types
• Creating snapshots and deleting disks
• Deleting virtual machine images
• Deleting unused disks allocated to virtual machines
• Deleting unused IP addresses.

The day they started using Rabbit, Bitrise was able to identify an opportunity to save $2000 per month on their cloud costs. At the time, Bitrise was in the process of a major architectural change and migration. During the migration process, compute engines were manually deployed from the Google Cloud Console and were not cleaned up after. The missing cleanup left the virtual machines (VMs) in an idle state. Using our tool, Bitrise discovered this wasteful practice and was able to save $2000 on a monthly basis.

Bitrise quickly understood that with Rabbit’s ability to provide an organizational and service-level overview of their company’s costs, they could make their operations more efficient. It is one of Rabbit’s most important goals to bring cost transparency. As evidenced by the example above, this can result in significant cost savings for any company.

Optimizing BigQuery use

Our close partnership with Bitrise enabled us to use the company’s feedback to further improve Rabbit’s functionality. A key example is the BigQuery dashboard to analyze daily data usage per queries, tables, and accounts.

Using amazing feedback from Bitrise’s Staff Data Engineer Balázs Francsics, we quickly tuned Rabbit to be a truly useful tool for any data engineer working with BigQuery. Besides cost optimization, Rabbit also provides actionable insights to improve performance by adding recommendations to tables for partitioning or clustering.

A key finding for Balázs was the cost of an hourly query that generated a user report for sales purposes. Generating this report cost Bitrise thousands of dollars each month. When inquiring if that frequency was justified, Bitrise found that the report was only used once a week during the weekly sales meeting. Running the query once every Friday was sufficient, enabling Balázs to reduce the costs of this query by 1/168.

Table column statistics helped Balázs further understand and optimize the company’s BigQuery tables. And the Accounts feature was valued for its ability to notify people querying a certain table to be cautious about a specific table update or when adding new columns.

Committed Use Discounts

Committed use discounts (CUD) can be challenging to understand and calculate. Most GCP clients resort to extrapolating from historical data or just guessing a commit. Understanding that and realizing its importance, the Rabbit team built a strategic tool for GCP users specifically targeted at CUD.

By combining all the dimensions needed to calculate the optimum for cloud usage, Rabbit users no longer need to spend time processing long and detailed documentation about committed use discounts and executing their own CUD calculations. Instead, they can quickly access Rabbit’s recommendations and focus on building new functionality.

The case of Bitrise is a great example. The company has an extensive Compute Engine operation and wanted to leverage the discounts provided by Google if resources are reserved for 1 or 3 years. By using Rabbit’s CUD feature to fine-tune the savings they were planning based on their roadmap, Bitrise managed to save 20% on their annual spending.

Before getting started with Rabbit, Bitrise’s team had to spend weeks poring over a manually updated spreadsheet, only to realize it quickly became unmanageable and confusing. In contrast, Rabbit provides a clear interface with three convenient options:

• Smaller savings for stable usage
• Maximized savings for cost control
• An easy-to-use and customizable calculator

The chart above is a resource-based commitment suggestion for a machine type used by Bitrise. Rabbit’s CUD planner knows that it can be worth committing over the minimal usage: even if Bitrise doesn’t utilize the suggested commitment on the weekends, this loss is more than made up during weekdays.

Sign up for a free trial, or contact us to try Rabbit on demo data.