Understanding data storage architectures

Time Series Databases (TSDBs) are ideally suited to handle metric data, as metrics need to record data at specific points in time, and TSDBs are structured to make this data easy to record and query. There are several TSDBs available, but as this book is focused on Grafana, we will only discuss Graphite, Prometheus, and Mimir in this section. This is aimed at giving you an understanding of the structure of data as it is stored, as well as an overview of how Mimir allows organizations to scale their data beyond the capabilities of Graphite and Prometheus.

Graphite architecture

Graphite has several components; we will discuss the storage component Whisper here. The Whisper TSDB uses a flat file structure, where each unique time series is a fixed-size file. This size is determined by the configuration of resolution and retention configured in Whisper. Gathering this data for a search requires each of these files to be read, which quickly...