Case Study: Philips Wellcentive restores Hadoop performance and meets SLAs with Pepperdata Platform Spotlight

Pepperdata Shines a Light Into Philip's Wellcentive Big Data Cluster

Philips Wellcentive, a cloud-based population health management and data analytics company, was struggling with a critical performance issue in its Hadoop environment. Its nightly eight-hour MapReduce job, which processes medical records for more than 35 million patients, began failing due to unpredictable cluster behavior, missed SLAs, and lost revenue, and existing monitoring tools could not pinpoint the root cause. Pepperdata’s Platform Spotlight was used to gain deeper visibility into CPU, memory, disk I/O, and network usage.

With Pepperdata Platform Spotlight, Philips Wellcentive quickly identified disk I/O contention caused by the merge phase on a mapper and large temporary data volumes from its “slash and burn” process. The team tuned spilled records from 8x to 2x, improved cluster throughput with Capacity Optimizer, and resolved the issue within days of installation. As a result, jobs completed on time again, SLAs were met, and the team could refocus on core initiatives.

Philips Wellcentive

Geovanie Marquez

Software Architect