Immediately query events after they occur and unify real-time and historical insights. Aggregate, drill-down, and slice-n-dice N-dimensional data with consistent, sub-second response times.
Scalable and available; consistent and predictable performance in multi-tenant environments. Existing clusters have scaled to petabytes of data and trillions of events, ingesting over 1M events/second.
Druid is built for OLAP. It excels at aggregations and supports a variety of filters and calculations. Leverage approximate algorithms to improve performance, and use exact algorithms when correctness matters.