Case Study: Grab achieves real-time observability at scale with StarTree Apache Pinot

Grab powers real-time observability with StarTree in 1 second

Grab, a Southeast Asian super app company offering ride-hailing and other services, faced a challenge with fragmented observability data. Their metrics were scattered across multiple systems, requiring manual queries and tribal knowledge to access. To unify this data into a single, real-time platform, Grab partnered with vendor StarTree and implemented a solution built on Apache Pinot.

StarTree's Apache Pinot formed the core of Grab's new metrics platform, Midas. This system provided a centralized source for reusable metrics, allowing for flexible, real-time analysis without pre-aggregation. The solution achieved massive scale and performance, handling over 10 million monthly requests with a p95 query latency of one second. This represented a 90% growth in usage over six months, providing consistent and actionable observability for operations teams and data scientists.

Grab