Based on this Udemy course
- Prometheus uses a dimensional data model
- A set of labels identifies the metric
| Metric Name (Dimension) | Label | Sample |
|---|---|---|
| Temperature | location=outside | 90 |
- Prometheus includes a flexible query language
- The metrics are stored in memory and local disk in its own efficient format
- It is written in Go
- Prometheus collects metrics from monitored targets by scraping metrics HTTP endpoints
- All data is stored as a time series
- Every time series is identified by metric name and labels (key value pairs)
- The time series data also consists of samples which can be float64 or ms time stamp
- Notation:
<metric-name>{<key>=<value>, <key2>=<value2>,...} - GQL used