MetricStash

MetricStash: Open-Source Time Series Databases

MetricStash is an open-source time series database optimized for high-ingestion workloads. It is designed to efficiently handle high volumes of numeric data with automatic aggregation, downsampling, and data expiration. MetricStash is useful for DevOps, IoT, and monitoring use cases.

What is MetricStash?

MetricStash is an open-source time series database built specifically for high-velocity numeric data such as metrics, events, and analytics. It is optimized for high-ingestion workloads common in modern monitoring pipelines, providing automatic rollups, retention policies, and data downsampling. Key capabilities include:

• High ingestion rates exceeding 1 million metrics per second per node
• Automatic aggregation of raw data into rollup tables for different time ranges
• Flexible retention policies to control storage consumption
• Tag-based metadata for slicing and dicing metrics
• Native support for Graphite data retrieval APIs
• Horizontal scalability through sharding

MetricStash is implemented in Go and leverages Facebook's Gorilla TSDB paper for its core storage format. It delivers exceptional write performance and compression ratios. The project is open source under the Apache license.

Typical use cases for MetricStash include:

• Infrastructure monitoring platforms
• Log collection and analytics
• IoT and sensor data
• DevOps analytics
• Cloud cost management
• Application performance monitoring

MetricStash Features

Official Links