square/cubeArchived
Cube
Cube is a time-series analytics platform and event data store designed for real-time performance monitoring. It functions as a metrics engine that ingests timestamped event streams and persists raw logs to enable the computation of statistical summaries, quantiles, and histograms.
The system distinguishes itself through a reactive processing model that automatically invalidates metric caches when new events arrive, ensuring query results remain current. It supports both real-time event streaming via persistent connections and the calculation of post hoc statistics from stored event sets.
The platform's capabilities include time-series data analysis, event data aggregation, and the hosting of web-based analytics dashboards by serving static assets from the local filesystem.
Features
- Event-Based Metric Computations - Derives aggregate statistics and histograms by applying mathematical expressions to sets of stored event data.
- Real-time Performance Monitoring - Provides a platform for collecting and analyzing live event streams to track system behavior and performance trends.
- Statistical Summaries - Converts raw event streams into aggregate statistics, quantiles, and histograms for high-level system observation.
- Event Data Ingestion - Accepts event data via network protocols and persists them for later metric computation.
- Event Timestamp Definitions - Stores raw timestamped events to enable incremental processing and subsequent aggregate computation.
- Event Stores - Persists raw timestamped logs to enable historical analysis and reactive metric computation.
- Raw Event Persistence - Stores timestamped events in their original form to allow for post hoc statistical analysis.
- Post Hoc Statistical Analysis - Derives aggregate statistics, quantiles, and histograms from stored event sets after data collection is complete.
- Reactive Processing Engines - Aggregates event data into metrics and manages cache invalidation to provide up-to-date query results.
- Time Series Analysis - Stores and queries timestamped event logs to compute statistical summaries and historical performance metrics.
- Time-Series Analytics Platforms - Ingests event streams and computes statistical summaries, quantiles, and histograms for performance monitoring.
- Time-Series Aggregations - Processes raw event streams into statistical summaries using mathematical expressions and time-based intervals.
- Statistical Metric Calculators - Calculates medians and boundary values (quantiles) across collected event datasets.
- Event-Driven Metric Updates - Updates metrics reactively by applying expressions to event data as it arrives.
- Analytical Event Logs - Stores raw timestamped events to enable post hoc statistical analysis and quantile calculations.
- Reactive Metric Invalidators - Automatically clears cached metric values upon new data arrival to ensure query results remain current.
- Analytics Dashboards - Hosts web-based interfaces for visualizing and exploring analytical data and system behavior patterns.
- Cache Clearing - Provides mechanisms for clearing cached metric values to ensure evaluators use current data.
- Event Retrieval Services - Fetches specific timestamped events over a defined time range via network requests.
- Log Aggregator Query Metrics - Transforms event data into structured metrics using query language expressions and defined time resolutions.
- Real-Time Data Streams - Pushes real-time event data to clients through a persistent network connection.
- Real-time Event Streams - Pushes new events to clients continuously as they arrive through a persistent connection.
- Stream Piping - Implements data stream piping to push real-time event data to clients with configurable delays.
- Analytical Dashboards - Hosts real-time dashboards for slice-and-dice and drill-down operations to display system trends.
