How Nanit drives engineering efficiency and innovation with Redpanda

Challenge

RabbitMQ’s limitations stall innovation and collaboration

With an award-winning baby monitor and sound machine, Nanit is creating an entirely new experience for parents with innovative technology and personalized, science-backed insights, supporting every stage of a growing family and empowering parents like never before.

As Nanit’s popularity grew, so did the demands on its backend infrastructure. Real-time event data was at the heart of all operations, powering notifications, enabling historical analysis, and driving product innovation. At the time, Nanit relied on RabbitMQ to drive its micro services architecture. Each service—whether analyzing baby sleep data, sending alerts, or storing video footage—needed to send and receive data efficiently, and RabbitMQ facilitated that communication.

However, one of RabbitMQ’s biggest limitations was its designed purpose as a message broker, not an event store. It was impossible to replay historical event logs, so Nanit couldn’t reuse data for analytics, debug issues, or build features based on historical insights.

“To create a new feature, we needed data from multiple sources,” said Senior Backend Developer, Eddie Mishelevich. “That meant fetching data from many different APIs and storage systems. Some features just weren’t worth the effort to implement, so we had to move on."

With ever-increasing event volumes, RabbitMQ struggled with high throughput, leading to queue bottlenecks, dropped messages, and reliability issues.

Meanwhile, Nanit was moving from a chapter-based structure to a pod-based model, where each pod (or team) operated independently with full-stack capabilities. It was a major architectural shift, and Nanit needed a way for services to publish event streams that other pods could access freely and quickly. Except that teams building features had limited data to work with, since RabbitMQ only provides new events, not historical events.

More critically, without a partner managing their streaming data solution, Nanit’s engineering teams spent much of their time troubleshooting failures, queue build-ups, and message drops — all of which slowed development.

Nanit needed a dedicated event streaming platform that could scale with their growing real-time data needs, support event persistence and replay, and provide seamless, low-latency access to event data across teams. It was the only way they could continue to innovate and deliver value to new parents.

“We needed a way for services to expose a public stream of events so other pods could consume it freely, without manual coordination or API integrations.”

Why Redpanda

A fully managed, high-performance streaming solution

Nanit evaluated three options: self-managing their clusters, using Amazon Managed Streaming for Apache Kafka (Amazon MSK), or adopting a fully managed event streaming solution.

Ultimately, Nanit chose Redpanda Cloud on Amazon Web Services (AWS), a fully managed Kafka-compatible solution that offered scalability, performance, and reliability without any operational burden.

Redpanda offered several benefits that other platforms lacked:

• Guaranteed data safety with zero loss. Redpanda’s architecture guarantees persistent, fault-tolerant event logs and data integrity across its real-time event pipeline.
• A hassle-free, fully managed solution. Nanit offloads maintenance, upgrades, and infrastructure issues to the Redpanda team, freeing their engineers to focus on product development.
• Seamless developer collaboration. Redpanda enables teams across Nanit to easily access, consume, and share event data without manual coordination or custom API workarounds.
• Simplified infrastructure management. Redpanda’s proactive monitoring identifies issues early—often before Nanit’s internal monitoring tools detect them.
• Effortless scalability. Redpanda allows Nanit to seamlessly increase event streaming capacity as their platform grows, without downtime, operational complexity, or high costs.
• Tiered storage. Nanit offloads older event data into cost-effective object storage. They can still retain large and instantly accessible volumes of data without expensive storage costs.

Beyond Redpanda’s technology, Nanit found a true extension of their team. “Redpanda’s team is incredibly responsive, professional, and proactive,” Eddie said. “They alert us about issues in our cluster or that we’re running out of resources, and they’ll often mitigate it without asking.”

Results

Faster development, greater reliability, and effortless scaling

Since migrating to Redpanda Cloud on AWS, Nanit has modernized its data architecture, transforming from a basic queue-based system to a sophisticated event sourcing platform. This architectural evolution has allowed teams to think about data as persistent, reusable streams rather than ephemeral messages, fundamentally changing how they design features and systems.

No longer over-reliant on specific APIs, Nanit’s architecture is more resistant and flexible than ever before. Redpanda’s zero-data-loss architecture ensures that every event is stored, accessible, and reusable, which enables Nanit to build more intelligent features powered by real-time insights. And for the first time, engineers can easily access and share event data without complex integrations or manual coordination.

“Developers don’t need to work harder to expose their data, and teams wanting to consume that data can simply do so. This improves velocity, leads to faster feature development and deployment cycles, all of which help Nanit run a better business,” Eddie said.

As a result, Nanit was able to unlock completely new use cases, including:

• Streamlined event sourcing - Services can consume business events directly from Redpanda to build their own representation of state, increasing reliability and reducing dependency between services.
• Unified data to determine user eligibility - Instead of talking to multiple microservices to gather all the data needed to determine whether a user is eligible for a specific feature, the team can now push all the data into a Redpanda topic for an all-in-one view.
• Built-in simplicity - The integrated Redpanda Connect pushes data from topics to Parquet files on Amazon S3, skipping the need to maintain Kafka Connect separately.

Having a fully managed service also eliminates the operational burden of managing clusters. Engineers don’t waste time troubleshooting queues or managing disk space because Redpanda handles it all — saving them time and energy they can instead devote to higher-value work.

As Nanit continues to scale, Redpanda scales with them. The team recently doubled its cluster capacity without any downtime, disruption, or operational headaches. Redpanda has given Nanit’s team the peace of mind they need to expand the world of smart baby monitoring for the latest generation of parents.