Harness: Modern BI meets modern software delivery to revolutionize the developer experience
About Harness
Harness, the Modern Software Delivery Platform®, provides developer tools for delivering reliable code faster as well as dashboards with KPIs measuring developers' efficiency and experience across software lifecycles, such as deployment frequency, lead times, failure rates, and cloud costs. In 2023, Forbes added Harness to its Cloud 100 list, and Techstrong Group rated Harness' platform as the Best End-to-End DevOps Tool/Service.
Tell us your challenge. We're here to help.
Contact usLooker empowers Harness to streamline the developer experience and help their customers, as well as internal teams, achieve outcomes they didn't know were possible.
Google Cloud results
- Reduces projected development time by 50%
- Increases developer efficiency by 75%
- Simplifies data security, reliability, and management
- Provides the foundation for an AI-powered BI development assistant
Expedites customers' queries from minutes to seconds
As a startup in 2017, Harness made the strategic decision to include a homegrown business intelligence (BI) solution in its first product, Continuous Delivery & GitOps. By doing so, Harness enabled their customers to reduce developer toil, while also being able to measure efficiency, costs, and risks. The solution was a smash hit among global developers straightaway, and its popularity generated momentum for building a broader platform for other products.
The second product Harness developed was Cloud Cost Management, which gives organizations insights and controls over their multicloud usage. Because Cloud Cost Management would be dashboard-heavy, Harness decided to embed a third-party BI tool into its platform rather than evolve its custom-built BI software. "Our in-house dashboarding solution worked, but building that out to support future products and evolving it to meet customers' needs would require significant engineering bandwidth and maintenance costs," says Rohit Reddy, Senior Product Manager at Harness. "We'd basically be building and managing a parallel solution for custom dashboards."
"The semantic layer we could build with LookML would deliver the best ROI given the scale we anticipated. It was a strategic decision that has paid dividends for us."
—Rohit Reddy, Senior Product Manager, HarnessHarness searched for a customizable and scalable third-party BI tool that supported white-labeling and the diverse use cases of its customer base. Since Harness customers include a wide range of verticals and sizes, it was vital to choose a solution that would be flexible to meet their needs. Ease of maintenance was also a key concern, so Harness wanted a BI product that featured the same code base for SaaS and on-premises deployments. To simplify management, improve security and reliability, and streamline all development, Harness also wanted to build a unified semantic data layer with its BI solution so it could establish a standardized data model.
Looker checked all the boxes, and more. Reddy explains, "At the time, Looker seemed like more than we needed, but it addressed all our pain points and the semantic layer we could build with LookML would deliver the best ROI given the scale we anticipated. It was a strategic decision that has paid dividends for us." As a Google Cloud shop, Harness would also continue to use BigQuery for Cloud Cost Management and explore its use in future products.
Harness cuts development time by 50% with Looker
Harness used Looker to build an initial semantic data layer to ensure BI accuracy, consistency, and access controls. The layer specifies which user roles can access which data in the underlying BigQuery, as well as the data dimensions and formulas for calculating KPIs. Harness also used Looker to create dashboards for their Cloud Cost Management product that give both its customers' developers and executives relevant, real-time insights into cloud costs. Customers can also explore the governed data and calculations supporting the KPIs, and create their own reports on demand. Similarly, Harness' customer support teams leverage Looker to quickly customize dashboards and configure alerts for customers when needed.
"We saved at least three months of engineering development effort by embedding Looker in our platform rather than writing our own BI solution. And we enable a self-service BI approach that just wasn't possible with our homegrown solution."
—Rohit Reddy, Senior Product Manager, HarnessIn just three months, Harness built its Cloud Cost Management product. "We saved at least three months of engineering development effort by embedding Looker in our platform rather than writing our own BI solution," says Reddy. "And we enable a self-service BI approach that just wasn't possible with our homegrown solution."
What's good for the goose is good for the gander
Harness proceeded to develop 11 more software products, nine of which use Looker to give both customers and internal teams the real-time insights they seek. "Because Looker is embedded in our platform, we use it to create internal facing dashboards that provide platform-wide analytics," says Joseph Lopez, Staff Business Intelligence Engineer at Harness. "We use the Looker embed client within the custom dashboards section of our platform to view account-level insights. And with Looker, we can even see the ROI of the investments we make in pipelines and the time it takes to run them."
IT staff efficiency sees 75% surge
To better meet quickly evolving requirements, Harness recently updated the architecture of its semantic data layer. Rather than replicating and managing a data layer for individual production clusters, developers spent a month rewriting its data layer so it follows a hub-and-spoke model. As a result, developers now manage data in one central hub. "Today, we edit all our LookML code in one place and that propagates down to all the spokes," says Lopez. "By using this approach, we onboard customers faster and improve security, reliability, and management. We also increase efficiency by about 75% because developers no longer have to update four Explore files and test each one separately."
Accelerating customer queries from minutes to seconds
When the tables of one of its large customers grew to hold multiple terabytes of data, Harness built aggregate tables in Looker that include only the data needed to process queries supporting specific KPIs. By doing so, Harness decreased the amount of data analyzed in queries to 0.096%, which is 1,000x less than before. This sped up insights for its customers and reduced its own data processing costs. "Within a month, we spun up aggregate tables for one of our largest customers that satisfied a majority of their use cases and decreased their dashboard load times from tens of minutes to less than 10 seconds," says Lopez.
Harness now uses aggregate tables as a best practice for customers with large tables. "Looker aggregate tables are our under-the-hood UX magic that scale to support the ever evolving query needs of our customers," Lopez says. "By dynamically routing to the smallest available table, they enjoy faster queries and we save money by not having to change existing reporting to a different Explore."
BI is a differentiator, not a deterrent
Harness expanded its target customers and revenue opportunities by offering different tiers of BI capabilities. Enterprise customers who have the custom dashboards feature in their plan have access to out-of-the-box Looker dashboards that deliver immediate insights into their core KPIs. These enterprise customers can also use embedded Looker to create their own custom dashboards and visualizations. They can clone existing dashboards and then customize them by filtering out information, adding different KPIs, or changing how data is visually depicted. These enterprise customers can also use Looker to create custom reports and alerts, and deliver custom insights via Slack and email on demand and automatically.
AI that's beyond the buzzword
Harness is using AI and ML technologies in Google Cloud, such as Vertex AI to automate core tasks like code deployments, vulnerability remediation, as well as service deactivation and activation. Customers monitor these automated processes from their Looker dashboards. Harness also uses AI with Looker to deliver predictive insights, including projected efficiency scores for future cost model options. Just by delivering these types of automated controls and predictive insights in Cloud Cost Management, Harness customers reduce their cloud expenses by up to 70%.
"By fully exploiting the capabilities in Looker and our semantic data layer, we're continuing to accelerate the development of disruptive solutions like AIDA that drive growth and efficiency for our customers — and for Harness."
—Joseph Lopez, Staff AI Engineer, HarnessOne of Harness's newest innovations is a service called Artificial Intelligence Developers Assistant (AIDATM), which can automatically handle basic tasks for all user personas, saving significant time and money. With directions in natural language, developers can tell AIDA to automatically detect vulnerabilities and issues in builds and deployments. And after reviewing what they are, developers can direct AIDA to automatically implement suggested fixes.
AIDA can also automatically answer questions and create custom dashboards by using Looker. The semantic layer ensures AI algorithms and users only see the data that's available to them, in formats that are accurate and consistent. "Executives and even DevOps engineers don't usually have in-depth BI development experience to feel comfortable creating their own dashboards," says Lopez. "AIDA alleviates any initial learning curves by showing them what's possible with Looker."
A Cloud Cost Management user can ask AIDA, "What's my total AWS cost?," and AIDA will automatically run the query and tell them. In just seconds, AIDA can also use Looker to generate an explorable chart that answers a more sophisticated question by a user like, "What are my total Google Cloud costs over the past 12 months?"
"Our use of Looker is a key differentiator for us," says Lopez. "The amount of data we expose to our customers in dashboards and the BI capabilities we give them to generate visualizations are unmatched. By fully exploiting the capabilities in Looker and our semantic data layer, we're continuing to accelerate the development of disruptive solutions like AIDA that drive growth and efficiency for our customers — and for Harness."
Tell us your challenge. We're here to help.
Contact us
About Harness
Harness, the Modern Software Delivery Platform®, provides developer tools for delivering reliable code faster as well as dashboards with KPIs measuring developers' efficiency and experience across software lifecycles, such as deployment frequency, lead times, failure rates, and cloud costs. In 2023, Forbes added Harness to its Cloud 100 list, and Techstrong Group rated Harness' platform as the Best End-to-End DevOps Tool/Service.