Observability 📊

End-to-end observability, distributed telemetry, and site reliability engineering (SRE) for modern cloud-native systems.

Start here

• Observability fundamentals — The signals, what each is good for, and how they fit together
• Instrumenting applications — What to measure (RED and USE), naming, labels and histogram buckets
• Alerting — Symptoms over causes, burn-rate alerts, Alertmanager routing, reducing noise

1. OpenTelemetry — the telemetry standard

Comprehensive 13-part curriculum covering the OTel architecture and data model:

• 🏛️ OpenTelemetry Overview & Architecture — Core specification, SDKs, and data model
• 🧭 Signals Overview — Traces, Metrics, Logs, and Baggage
• 🛰️ OTel Collector Architecture — Receivers, Processors, Exporters, and Extensions
• 🌐 OTLP Protocol — Protobuf over gRPC (port 4317) & HTTP (port 4318)
• 🔀 Context Propagation — W3C traceparent, tracestate, and Baggage injection/extraction
• 🏷️ Semantic Conventions — Standard attributes for HTTP, DB, RPC, and Cloud
• 🔍 Distributed Tracing 101 — Spans, Tracers, SpanProcessors, and ParentBased Samplers
• 📈 Metrics 101 — Counters, UpDownCounters, Gauges, Histograms, and Temporality
• 📜 Structured Logging 101 — LogRecord data model, OTel Bridge APIs, and correlation
• ☸️ Kubernetes Observability — OpenTelemetry Operator, Target Allocator, and DaemonSets
• 🛠️ End-to-End Implementation Lab — Hands-on multi-service instrumented lab
• 📖 OpenTelemetry Glossary — Quick reference for key concepts

2. Metrics

• Prometheus architecture — Pull model, series and labels, metric types, the TSDB, scaling with remote write
• PromQL — Selectors, rate, aggregation, histograms, recording rules
• Instrumenting applications — Client libraries in Go and Python, and the mistakes that make metrics expensive

3. Logs

• Logging — Structured logs, levels, the collection pipeline, Loki and LogQL, cost control

4. Traces

• Distributed tracing — Reading a trace, context propagation, head and tail sampling, span metrics, backends

5. Platforms and visualisation

• Grafana — Dashboard design, variables, dashboards as code, correlating signals
• The LGTM stack — Loki, Grafana, Tempo and Mimir as one platform: architecture, deployment modes, run or buy
• eBPF for observability — Kernel-level visibility without instrumentation, and its limits

6. Acting on it

• Alerting — What deserves a page
• SLOs and error budgets — Turning telemetry into reliability targets
• Incident management and on-call

Related Topics

• ☸️ Kubernetes Operations & Observability — Kubelet metrics, cAdvisor, and cluster triage
• ☸️ Observability on EKS — Control plane logs, Fluent Bit, ADOT, Managed Prometheus, OpenSearch
• ☁️ AWS monitoring — CloudWatch metrics, logs, alarms and dashboards
• 🔷 Azure Container Insights — Managed Prometheus and Azure Monitor
• 🐧 Linux observability — journalctl, strace and host monitoring
• 🔐 SIEM — The security use of the same telemetry