Overview
A comprehensive time series service designed to handle multiple data types and query patterns for event-driven applications. The service supports both aggregated metrics (counting and histograms) and detailed event retrieval with flexible filtering capabilities.
Core Functionality
Data Types Supported
The service handles two primary data categories:
- Aggregated Metrics: Pre-computed counts and histogram distributions for efficient querying of statistical data
- Raw Events: Individual event records with full metadata preservation for detailed analysis and audit trails
Key Features
Metrics Storage and Retrieval The service maintains counting metrics that track occurrences of specific events over time, allowing for rapid retrieval of volume trends. Histogram functionality captures value distributions, enabling percentile calculations and statistical analysis of event characteristics like response times, file sizes, or user activity levels.
Event Filtering and Querying Events can be retrieved using multi-dimensional filtering based on object identifiers (such as user IDs, device IDs, or resource identifiers), event types (login, purchase, error, etc.), and specific event values or properties. Time-based queries support flexible time ranges from seconds to years, with configurable precision.
Temporal Flexibility The service supports various time window queries including fixed periods, rolling windows, and custom date ranges. This enables both real-time monitoring and historical analysis use cases.
Use Cases
This architecture serves applications requiring both high-level operational visibility through metrics dashboards and deep-dive investigative capabilities through event-level data access. Common applications include system monitoring, user behavior analytics, financial transaction tracking, and compliance auditing.
Architecture Benefits
By supporting both aggregated and granular data access patterns, the service optimizes for different query performance requirements while maintaining data fidelity. Fast metric queries enable real-time dashboards, while detailed event retrieval supports forensic analysis and detailed reporting needs.