Observability explained: metrics, logs, traces, and the Prometheus-Loki-Grafana stack

Modern distributed systems — running across dozens of services and containers — demand more than traditional alert-based monitoring, which only catches problems anticipated in advance. Observability is the property of a system that lets engineers answer unexpected questions using externally exposed data, without deploying new instrumentation each time. It relies on three data pillars: metrics (numeric time-series values ideal for alerts and dashboards), logs (timestamped text records carrying rich event context), and traces (end-to-end records of a single request's journey across multiple services). Tools like Prometheus, Loki, and Grafana form a practical open-source stack that addresses all three pillars, with Loki deliberately indexing only a small set of labels to keep log storage costs manageable. This article is the first in a series that will build a fully functional observability stack using real Docker Compose examples.