Case Study: Parse.ly Detects Performance Problems Early with Metricly

How Parse.ly Leverages Metricly to Detect Performance Problems Early

Parse.ly, a real-time analytics platform processing tens of thousands of requests per second and more than 60 billion requests per month, needed a better way to spot performance issues across its data pipeline. Its traffic naturally fluctuates with news, holidays, customer growth, and other outside factors, making threshold-based monitoring ineffective. Parse.ly turned to Metricly’s anomaly detection and custom metrics monitoring to keep up with its rapidly changing workload.

Using Metricly’s adaptive monitoring, Parse.ly ingested application metrics via StatsD and gained real-time anomaly detection, utilization tracking, and AWS cost analysis. Metricly helped the team detect deviations and performance degradations much earlier than before, sometimes uncovering issues they would not have found otherwise. As a result, Parse.ly improved capacity planning, balanced cost and performance more effectively, and protected the quality of its users’ experience.

Parse.ly

Chris Clarke

DevOps Lead