Achieving over 99.99% Resiliency Using AWS Services with Sprinklr

Founded in 2009, Sprinklr provides a customer experience solution to help businesses gain insights and track marketing through social media. The company has expanded over the years to help engage customers across more than 30 digital channels. It now offers four robust product suites that provide various customer experience and marketing tools, such as marketing analytics and social listening. In 2022, the company launched the Sprinklr Service product suite. When building this CCaaS solution, Sprinklr quickly realized that its customer base required increased resiliency.

To take full advantage of Sprinklr Service, customers needed high availability—and expected it. Although Sprinklr already had a resiliency of 99.9 percent, which worked well before these new offerings were implemented, that did not meet its new service-level agreement obligations of more than 99.99 percent resiliency. The company was born on AWS but needed to change its architecture, both on AWS and internally, to meet these new expectations.

One of the internal changes was the implementation of horizontal and vertical auto scaling to handle changing demands. In addition to that, Sprinklr implemented key resiliency patterns, like circuit breaker, timeout, bulkhead, and rate limiters, across all microservices to handle faults gracefully and prevent system overloads. Sprinklr also reduced the blast radius of backend services in terms of exposure to other components to minimize the scope of impact. One major point of rearchitecting that Sprinklr made on AWS was to migrate key value use cases to Amazon Keyspaces (for Apache Cassandra)—which is a scalable, highly available, and managed Apache Cassandra–compatible database service—to achieve scalability and resiliency. Sprinklr also uses Amazon ElastiCache—which offers near real-time performance for near real-time applications—to enhance performance.

Over 2 years, Sprinklr enhanced its solution using a combination of architectural best practices and managed AWS services. This step-by-step rollout facilitated a 99.99 percent uptime for critical workloads and services with virtually no interruptions for customers. “We were confident that if we moved to managed AWS services and made these architectural changes, we’d achieve the resiliency we needed,” says Abhay Bansal, senior vice president of engineering at Sprinklr.

Sprinklr uses a variety of managed AWS services, including Amazon Elastic Kubernetes Service (Amazon EKS)—which is the most trusted way to start, run, and scale Kubernetes—to manage its Kubernetes workloads. “Migrating to Amazon EKS was pivotal for us,” says Nitin Goyal, vice president of engineering at Sprinklr. “We could scale up to abnormal workloads in essentially no time.” Sprinklr also uses Amazon Relational Database Service (Amazon RDS), which is an easy-to-manage relational database service optimized for total cost of ownership, and secures and stores its data on Amazon Simple Storage Service (Amazon S3), which provides object storage built to retrieve virtually any amount of data from anywhere.

The backbone of Sprinklr’s current architecture on AWS is Amazon Elastic Compute Cloud (Amazon EC2), which offers secure and resizable compute capacity for virtually any workload. Sprinklr has always been an early adopter of new Amazon EC2 instances, which has empowered it to make use of innovations and scalable compute capacity. For example, the company runs some of its Amazon ElastiCache workloads and 80 percent of its Amazon EKS and Amazon Keyspaces workloads on Amazon EC2 instances powered by latest AWS Graviton processors, which offer the best price performance for cloud workloads running on Amazon EC2, to enhance compute performance.

Sprinklr also uses Amazon EC2 I4i instances, which are powered by third-generation Intel Xeon Scalable processors, to boost its input/output performance. Using I4i instances, Sprinklr can run sustained peak MongoDB workloads with lower latencies and one-tenth the occurrences of downtime.

By changing backend infrastructure and using Amazon Keyspaces, Sprinklr reduced infrastructure degradations from three or four events per quarter to almost zero. “We reduced costs by 30 percent and still had 99.99 percent resiliency,” says Bansal. “Using Amazon Keyspaces, we achieved almost zero degradation with no operational overhead.”

Sprinklr has not only exceeded customer resiliency requirements but also reduced latency using AWS. Customer-observed tail latencies have decreased by over 80 percent. With the rearchitected infrastructure, Sprinklr’s customers have access to fast, highly available, and secure services.