GetGo, a Singapore-based car-sharing company, faced an observability challenge: How could it connect monitoring data across a growing, distributed environment? When problems occurred, engineers had to correlate information manually across services, while limited visibility into mobile user sessions made it harder to understand what customers were experiencing.
In this email interview, Kirivarnan Kumarasamy, GetGo’s Director of Engineering, explains why the company adopted Datadog, how it integrated observability across its technology stack and engineering teams, and where AI-assisted and agentic capabilities could fit into its future technology plans.
Why did GetGo choose Datadog for observability?
Since launching in 2021, GetGo has expanded to more than 3,000 vehicles across 1,700 locations and over 400,000 registered users. As the service grew, so did the complexity of our technology environment. We operate a fully cloud-native, event-driven microservices architecture on AWS, with native iOS and Android applications supporting our customer base.
From a business perspective, one of our priorities was improving our ability to understand and respond to issues experienced by users, particularly on mobile. We also wanted to reduce the engineering time spent on manual troubleshooting and bug investigations.
We selected Datadog because it combined these capabilities on one platform. Its support for our technology stack simplified implementation, while its application performance monitoring (APM), infrastructure monitoring, logs, and, in particular, Datadog’s Real User Monitoring (RUM, a tool for monitoring application performance and user activity in real time) addressed some of the largest gaps in our previous set-up.
RUM gives us visibility into mobile user sessions and behaviour that we previously couldn’t see, helping us pinpoint and resolve critical performance bottlenecks before they affect the customer experience.
What limitations did you encounter with your previous monitoring and observability tools?
We initially relied on Amazon CloudWatch, but as our platform and distributed architecture grew, it became increasingly difficult to get a complete picture of what was happening across our systems.
We lacked comprehensive APM capabilities and, more importantly, we didn’t have RUM for our mobile applications. This meant we had limited visibility into user sessions, application behaviour, and errors from the customer’s perspective.
Our monitoring was also fragmented, so when an issue occurred, engineers often had to correlate information manually across different services to understand what had happened and identify the root cause. Troubleshooting and root cause analysis could take several hours, adding to the team’s workload and taking engineers away from development.
How has Datadog improved GetGo’s engineering operations?
The biggest change has been the reduction in the time our engineers spend investigating issues. Bug investigations that previously took several hours now require roughly 55% less time. This allows our engineers to spend more time developing features rather than troubleshooting manually.
We’ve also seen a nearly 40% improvement in both mean time to detection and mean time to resolution. Monitoring and alerts allow our teams to identify and respond to issues earlier, helping us minimise downtime and maintain platform reliability.
RUM has given us visibility into mobile user sessions and application behaviour. This helps us identify problems earlier, including those that might previously have gone undetected.
More broadly, Datadog has become the backbone of our monitoring and reliability operations. All of our engineering teams use it for monitoring, troubleshooting, and improving system reliability. Our DevOps team uses it extensively for infrastructure monitoring and alerts.
How did GetGo implement Datadog across its technology stack and engineering teams?
One of the main considerations was integrating observability across a varied technology stack and a distributed microservices environment. We have back-end services running on Golang and C#, native mobile applications built with Swift and Kotlin, and web applications, all operating within our AWS environment.
Datadog’s multi-language SDK support made integration easier and, crucially, brought the different parts of our technology stack into a unified observability environment.
The onboarding experience was also very positive. The Datadog account and support teams provided hands-on guidance and recommended best practices throughout implementation, helping us roll out the platform.
From an organisational perspective, the biggest shift was getting teams accustomed to using a shared system for monitoring and troubleshooting. Today, Datadog is used across our entire engineering organisation, making monitoring and reliability part of how our teams work each day.
What are GetGo’s next engineering and observability priorities?
Our core rollout is already in place across the engineering organisation. We are now exploring whether capabilities such as Bits AI SRE and Cloud SIEM could improve log analysis, monitoring, and engineering productivity.
We are also examining how AI-assisted and agentic capabilities could improve engineering productivity and support more personalised experiences across our platform and mobile applications. The aim is to use more detailed data and automation to identify issues earlier, make better-informed decisions about performance and scaling, and give our teams more time for development.
The unified observability environment also supports our plans to scale. As we grow across the region, we want our engineers to work quickly while maintaining reliability and the customer experience.
