Victoria Metrics

Introduction

Victoria Metrics has been installed on the host Victoria in the Infra network witha an ip of x.x.x.132 to handle the long term storage of historical data for the Home Lab on Pear. The data itself will be collected by Prometheus on Pineapple. The data should be displayed on a dashboard on Grandilla. It is assumed that this will not become useful until the three hosts have been running for a while.

Unified Monitoring Stack

The Unified Monitoring Stack has now superseded the separate VMs that performed the Prometheus, Grafana and Victoria metrics roles and should be used instead. These pages will be left for reference but will not be updated so that all of the Performance monitoring will be undertaken by the Unified Monitoring Stack including the Victoria Metrics and storage of the historical data.

VictoriaMetrics: Our Scalable Time-Series Data Warehouse

In our home lab's observability stack, VictoriaMetrics serves as the robust and scalable long-term storage solution for all the time-series data collected by Prometheus. While Prometheus excels at collecting and querying recent metrics, its internal storage is not designed for indefinite retention or massive data volumes. VictoriaMetrics fills this crucial gap, providing efficient and durable storage for our historical monitoring data.

Architectural Placement and Core Function

VictoriaMetrics is deployed as a dedicated VM named Victoria. Its primary functions are:

• Remote Storage for Prometheus: It acts as the remote storage endpoint for Prometheus instances. In our setup, Prometheus on pineapple is configured to remotely write all collected metrics to victoria.
• Scalable Time-Series Database: VictoriaMetrics is specifically designed for time-series data, offering high write and read performance, efficient compression, and long-term data retention capabilities.

PromQL Compatibility: VictoriaMetrics supports PromQL (the Prometheus Query Language), allowing Grafana (and other tools) to query it directly, just like Prometheus. This ensures a seamless transition when accessing historical data.

Data Flow and Integration

• Prometheus Writes, Grafana Reads: Prometheus on pineapple continuously streams all collected metrics to VictoriaMetrics on victoria. When Grafana needs to display historical data, it still queries Prometheus, but Prometheus transparently fetches that data from VictoriaMetrics.
• Scalability: VictoriaMetrics allows us to retain months or even years of monitoring data, enabling long-term trend analysis and capacity planning without burdening Prometheus with storage overhead.

In essence, VictoriaMetrics provides the reliable, scalable, and long-term memory for our home lab's observability data, complementing Prometheus's real-time data collection and Grafana's visualization capabilities.