Learn modern observability with the LGTM Stack (Loki, Grafana, Tempo, Mimir) from the basics to production-grade: environment setup, history & the three pillars of observability, OpenTelemetry, Grafana fundamentals, metrics with Mimir & PromQL, logs with Loki & LogQL, log collection with Grafana Alloy, structured logging, distributed tracing with Tempo & TraceQL, OpenTelemetry Collector, correlation across pillars, alerting, SLI SLO SLA & error budgets, scaling Mimir/Loki/Tempo, multi-tenancy, Kubernetes observability, service mesh, CI/CD, continuous profiling with Pyroscope, and a production deployment checklist, in a total of 36 episodes.
Before touching Loki, Grafana, Tempo, and Mimir, you need to master the concepts of distributed systems, monitoring, and telemetry. In this episode you set up Docker, local Kubernetes, and supporting observability tools, then verify your first environment.

Observability evolved from traditional monitoring into a discipline that emphasizes the ability to ask any question about your system. This episode traces that evolution, the problems it solves, and where the LGTM Stack sits among ELK, PLG, and commercial solutions.

Metrics, logs, and traces are the three types of telemetry that form the foundation of observability. This episode discusses the concepts and philosophy behind each pillar, from metric types to context propagation, and how to correlate all three into a single investigation flow.

OpenTelemetry is the open-source standard for generating and collecting telemetry without being tied to a vendor. This episode discusses the history of the OpenTracing and OpenCensus merger, the API, SDK, and Collector components, and why OTel became the foundation of modern instrumentation.

This episode turns theory into practice: running Grafana, Loki, Tempo, and Mimir along with the OpenTelemetry Collector and Grafana Alloy with Docker Compose. You'll also connect data sources and send the first telemetry from a sample application.

Grafana is the visualization center of the entire LGTM Stack. This episode discusses server and plugin architecture, UI navigation, how to create your first dashboard and panels, using variables and templating, and the panel types and query editors available.

Mimir is a metrics backend that is Prometheus-compatible and horizontally scalable. This episode discusses its component architecture, a comparison with Prometheus, the remote write path, and metric types and the dangers of high cardinality.

PromQL is the query language for reading metrics from Mimir. This episode covers instant vectors and range vectors, operators and aggregation, common query patterns like rate and histogram_quantile, important functions, and best practices for query performance.

Metrics don't appear by themselves — applications must be instrumented. This episode covers the Prometheus and OpenTelemetry SDK instrumentation libraries, creating counters, gauges, and histograms, auto-instrumentation, and how to expose metrics via a metrics endpoint and OTLP.

Loki is a cost-effective log aggregator with an index-free approach. This episode discusses its component architecture, how label-based storage works, a comparison with Elasticsearch, and log parsing strategies from structured formats to label extraction.
