Free to read ยท Three levels
Observability: signals, semantics, and trustworthy evidence
Eighteen developed lessons cover metrics, logs, traces, OpenTelemetry, collection quality, identity, interpretation, sampling, privacy, and investigation.
Start learningIn this course
Learn to build evidence that answers concrete diagnostic questions instead of collecting signals without a purpose. A fictional parcel service explains event, acceptance, processing, and visibility boundaries, and how units, sampling, transformations, and query choices change the meaning of reported data.
Each level contains six developed lessons with explained examples, diagnostic mistakes, worked exercises, and official references. Python examples use local synthetic values only. Every YAML, JSON, and header block is non-executable teaching material; no SDK, collector, backend, or running dashboard is accessed.
What you will learn
- Choose metrics, events, and traces for explicit questions with truthful units, boundaries, and identity.
- Diagnose loss, delay, selection bias, aggregation errors, and display mismatches through clear contracts.
- Review context propagation, sampling, privacy, cost, and migrations while stating the limits of tested evidence.
Your learning path
Three levels
Basic: signals and measurement meaning
Understand questions, metric semantics, structured events, traces, distributions, and collection loss.
- Observability starts with questions
- Metric types, units, and temporality
- Structured logs and event semantics
- Traces, spans, and propagated context
- Histograms and honest latency summaries
- Collection pipelines and visible data loss
Intermediate: instrumentation and interpretation
Review business instrumentation, ordered transformations, identity, queries, dashboards, and quality-aware alerts.
- Instrumentation boundaries and ownership
- Collector transformations and routing contracts
- Resource identity and cardinality budgets
- Query populations and valid aggregation
- Dashboards as diagnostic interfaces
- Alerts and measurement quality
Advanced: bias, trust, and evidence engineering
Analyze sampling, async relationships, cost, privacy, migrations, and cross-signal investigation.
- Sampling and selection bias
- Asynchronous context and batch relationships
- Telemetry cost, retention, and pipeline capacity
- Privacy, baggage, and telemetry trust boundaries
- Telemetry contract tests and migrations
- Cross-signal investigation and evidence limits
Who this is for
- Developers, platform and operations engineers, and learners designing or interpreting telemetry.
- Teams seeking clear instrumentation contracts and useful signal correlation with cost and privacy controls.
Before you start
- Basic programming, requests, and service concepts; HTTP and JSON familiarity helps but no monitoring platform or cloud account is required.
- Basic Python reading is sufficient for fixtures, with no package, agent, or service installation.