Metrics are the cheapest signal and the one every alert is built on. This module builds a metrics library by hand, then uses the real one, then learns to query it without being fooled.
| # | Lesson | The question it answers |
|---|---|---|
| 01 | Metric Types and the Data Model | What is a time series, and what are counters, gauges and histograms? |
| 02 | Instrumenting a Go Service | What do I measure in a service, and how is it exposed? |
| 03 | PromQL Essentials | How do I turn raw counters into rates, ratios and percentiles? |
| 04 | Histograms | How are percentiles estimated, and how wrong can they be? |
| 05 | Cardinality | Why did the metrics bill explode, and how do I stop it? |
When you finish you can instrument a service with the RED metrics, expose them, query error ratios and p99 latency in PromQL, and predict how many time series your labels will create.