Freshworks reduces Snowflake optimization efforts from weeks to hours with Keebo
Tens of thousands of customers rely on the insights they get from Freshworks. With over 7 million queries run every day, costs for Snowflake-based data warehouses were growing at
an unsustainable rate. Even with a large data team (over 70 people), manual Snowflake optimization was not able to keep pace. After considering several alternatives, Freshworks turned to Keebo for fully-automated and dynamic optimizations, achieving significant direct cost savings and allowing the Freshworks team to add more users and features with confidence.
CHALLENGES
- Massive growth in analytics users and features
- Need to reduce costs without sacrificing quality or query response times
- Manual optimization could not keep pace at scale
RESULTS WITH KEEBO
- 10-15% monthly Snowflake savings even with increasing workloads
- Maintain strict SLAs on analytics performance
- Reduced burden on analytics team, reducing weeks of optimization each quarter to merely hours
Keebo’s automated optimizations always keep us on top of opportunities to save without impacting query performance.
Anand Chiddarwar
Principal Engineer, Analytics Platform
Savings Without Slowdown
With customers depending on them, Freshworks must maintain query performance SLAs. Keebo has several ways to control optimization, including an intuitive slider bar. For warehouses where performance is critical, Freshworks can set the slider bar so that optimizations will not impact query performance:
For warehouses where query performance is not critical, Freshworks can set the slider to
aggressively pursue cost savings, or leave it in “balanced” mode for Keebo’s patented machine learning technology to find the ideal optimization:
With Keebo, Freshworks is always in control, so that query performance is not impacted and
opportunities for savings are realized without effort from their data team.
Immediate Results
After a 30-minute setup, Keebo started showing results right away. After their trial concluded,
Freshworks rolled-out Keebo to Snowflake warehouses in four regions across the globe. Freshworks considered other alternatives as well, including continuing their manual optimization attempts with Snowflake’s native toolset. But none of these other methods could optimize dynamically around the clock quite like Keebo.
The Data Team’s Best Friend
The Analytics team at Freshworks considers what they do as a service to all of Freshwork’s
products. They know that customers are depending on the work of their diverse team of data engineers, data scientists, and machine learning engineers. Optimizing Snowflake was
simply taking too much time, adding up to weeks of effort each quarter. With the time this team saves thanks to Keebo’s optimizations, they can focus on adding new capabilities and more customers at a savings.
We offer analytics as a service. The cost and time savings we get from Keebo allows us to add more features and more customers confidently.
Kiran Kumar
Senior Staff Engineer, Analytics Platform
