Aggregations

CrateDB distributes queries across all nodes in the cluster, executing aggregation tasks in parallel. The result: real-time responses, even when data volume and velocity keep growing.

Key advantages:

• Run complex GROUP BY, AVG, COUNT, SUM, and MAX/MIN operations in milliseconds
• Aggregate time-series and event data in real time
• Scale linearly as you add nodes; performance grows with your cluster

CrateDB’s columnar storage format drastically reduces the amount of data scanned per query. Combined with vectorized execution and automatic indexing, this ensures aggregations are both fast and resource-efficient.

Why it matters:

• Read only the necessary columns per query
• Optimize CPU and I/O utilization
• Maintain sub-second response times, even at petabyte scale

CrateDB’s multi-model architecture allows you to aggregate across time-series, relational, JSON, text, and vector data — all within a single query.

Examples:

• Compute rolling averages and percentiles from IoT telemetry
• Aggregate JSON payloads from API events
• Combine numerical metrics with search filters or geospatial boundaries

With CrateDB, your dashboards always show what’s happening now. Data is continuously aggregated as it arrives, enabling up-to-the-second visibility for operations, performance tracking, or anomaly detection.

Typical use cases:

• Industrial IoT monitoring
• Infrastructure and log analytics
• Business KPIs and financial metrics

Traditional systems rely on materialized views or scheduled jobs to precompute aggregates. CrateDB eliminates that complexity, computing and updating aggregations on the fly, directly from live data streams.

Result: Instant, accurate, and always up-to-date metrics.

Traditional systems	CrateDB’s real-time engine
Depend on precomputed or cached results	Compute on fresh data instantly
Slow down with large joins or wide tables	Parallel, distributed aggregation
Manual tuning and index design required	Automatic indexing and optimization