Innovaccer uses a multi-cloud environment comprising AWS, Azure and GCP as cloud providers and Snowflake, Databricks and New Relic as SaaS developer tools. Being a data activation platform with more than 70 customers onboard and 1000 customer locations, their cloud usage is massive. They have over 50 AWS accounts, 15+ Azure accounts & several accounts at Snowflake and a dedicated infrastructure and FinOps team for their cloud operations.
Ankit Maheshwari, CTO and founding member of Innovaccer, has built the tech from the ground up and wanted a reliable DevOps Cloud product that can handle the size and complexity of their cloud infrastructure and ensure that all the stakeholders that, include non-technical job professionals, can make use of the product to help accelerate their MRR to cloud cost percentage goal. He trusted the capabilities of OpsLyft and chose us to help Innovaccer realize its business goals.
Innovaccer has a single tenancy infrastructure such that they have a separate cloud account for each customer, making it feasible to measure the cloud cost associated with each customer. Each customer has a dedicated customer engineering team comprising of several members, such as a customer engineering manager, product manager, and data analyst, among others, so they receive the highest level of customer satisfaction. However, this makes it a herculean task for the infrastructure team to handle the cloud cost tracking & optimization for these many accounts.
Before OpsLyft, Ankit Senapati, Director of Cloud FinOps, along with the finance and infrastructure team, used to manually monitor the MRR to Cloud cost percentage of each customer using Excel sheets. This meant they spent hours collating the data about each customer account from different cloud providers and converting them into business metrics. Not only did this result in a lot of manual human hours being consumed but also, it was challenging to maintain the order of prioritization.
To solve the major problem of observability at Innovaccer, Ankit Senapati took the help of a multi-cloud dashboard, budgeting, and role-based access management in OpsLyft’s product. Since OpsLyft’s product could collate all the cloud cost data from AWS, Azure, and Snowflake in a single dashboard, Ankit didn’t have to rely on Excel sheets anymore. In addition to providing infrastructure costs for each customer, OpsLyft was also integrated with the MRR data source through which custom unit cost analysis metrics such as % of MRR were set up for each customer engineering team.
With budgeting, Ankit could set cloud cost budgets with respect to MRR to cloud cost percentage target of each customer. Through role-based access management, he linked customer cloud accounts to each customer success engineer and product manager. He provided them with an individual personalized dashboard that showed them the cloud cost visibility only of the customer account they were working on. These dashboards provided them with the visibility of the MRR to the cloud cost percentage target, and they were able to understand if they were within the target or if they were going over it and needed to optimize the account.
User management and budgeting were not limited to the CS engineers and PMs. In fact, this facilitated organization-wide observability as Ankit Senapati could see and understand the overall cloud cost analysis of Innovaccer as he also got the data of the internal cloud accounts in addition to the customer cloud accounts. Ashish Singh, President of Platform and Tech Transformation, who leads the customer success engineering function at Innovaccer, could analyze the cloud cost performance of his department.
Ankit Maheshwari, CTO, was able to understand the forecast & projections of his cloud costs with respect to the business Metrics. This led to the leadership team making data-driven strategic decisions.
Naveen Gupta, CFO, and Mahesh Dutt, VP of Finance, used the dashboards to correlate the cloud cost of a customer to their business metrics & thus preemptively forecast the gross margins for the customers and figure out how they can accelerate their journey of reaching profitability.
After solving the massive issue of observability, the next focus was on deciding the order of prioritization. Each customer cloud account was monitored by the respective customer success engineer and product managers. If the MRR to cloud cost percentage was within the budget, they need not do anything. But if they were going over the budget, they need to perform cloud cost optimization to reach the month-end target.
However, there are more than 70 customer cloud accounts and it was nearly impossible for the infra team to optimize them on their own. To address this, if the budget of a particular customer account is being exhausted, the CS engineers and PMs could get cloud cost savings opportunities through OpsLyft’s cloud cost savings insights feature.
These cloud cost savings opportunities show them the list of underutilized resources that can be optimized in descending order which means the resources with maximum dollar savings are shown to them first. The criteria for these underutilized resources were a mixture of predefined rules from OpsLyft and user input from Innovaccer’s infra team.
Along with these underutilized resources, each resource has contextual information attached to it like the utilization status across multiple metrics, the application where the resource is being used & 3 sets of recommendations: First is cost optimized which means that they can replace the resource by an alternative that is significantly cheaper with lower performance and this is most suitable for dev and staging environments. Secondly, there is a balanced resource which is cheap but makes sure that the application performance doesn’t suffer significantly. Lastly, there is performance-oriented resource that provides maximum performance but is less costly than the current resource, making it the best option for the production environment.
After the CS engineers and PMs recognize the potential cost-saving insights, they create a Jira Ticket directly from the product and it is synced with Innovaccer’s internal Jira Board and it gets reflected there. Through this approach, the infrastructure team has a set of highly actionable cost-saving opportunities that can ensure that maximum savings are being made and they can achieve their MRR to cloud cost ratio goal faster.
Thus, Innovaccer’s cost optimization journey was completely automated with the help of Opslyft’s Cost Saving Insights.
After collaborating with OpsLyft, Innovaccer not only solved the issue of observability and prioritization order of cloud cost optimization, but they also achieved the following results: