ironSource exposed a Druid cluster for its external users, called Real Time Pivot, which runs at a scale of 2-3M events/sec, with tens of TBs added per day, while serving parallel queries within 1-2 seconds. We managed not only to achieve the goals of high throughput without getting backlogs and serving parallel queries with low latency, but we did all that at minimal hardware cost. In our presentation we’ll elaborate on how we achieved these three goals.
Keynote: Powering Event-Driven Data with Apache Druid
The distinction between OLTP and OLAP is becoming less relevant as data architectures shift toward entities and events. In this session, we’ll delve into how Apache Druid’s event-first approach synthesizes...
Watch nowClosing Keynote: Charting the Future of Druid
What lies ahead for Apache Druid? Join us as we explore the evolving landscape of Druid’s query and storage engines, and how they are positioned to address the biggest challenges in event data for the future. Speaker: Gian...
Watch nowSalesforce: Tracing Service Dependencies at Scale with Druid and Flink
At Salesforce, we manage approximately 300 million distributed spans to infer service dependencies. We have successfully utilized a combination of Druid and Flink to handle this scale with high availability....
Watch now