Observability is often reduced to dashboards and alerts, yet teams still struggle to explain system behavior during real incidents. This full-day masterclass treats observability as an engineering discipline rather than a tooling exercise.

Participants will design, implement, and analyze observability using OpenTelemetry and Grafana in a production-like environment. You will move beyond basic monitoring by applying intentional instrumentation, exploring semantic conventions, and evaluating trade-offs around cardinality, sampling, and cost. Through hands-on exercises, you will debug real failure scenarios using trace-first analysis and learn to correlate signals across layers to find root causes faster.

Full-Day Structure

• Introduction & Environment Setup: Baseline services, telemetry pipelines, and OpenTelemetry SDKs.
• Designing Observability Signals: Mapping system questions to metrics, traces, logs, and exemplars.
• Instrumentation with OpenTelemetry: Context propagation, span boundaries, and semantic conventions.
• Trace-First Debugging: Navigating critical paths, latency analysis, and signal correlation.
• Operational Trade-offs: Managing cardinality, sampling strategies, and cost vs. fidelity.
• Evolving Observability: Sustaining signal quality and scaling practices.

Key Learnings

• Question-Driven Design: How to design observability around meaningful system behavior questions.
• Signal Selection: When and why to use metrics, traces, logs, and exemplars.
• Effective Instrumentation: Hands-on service instrumentation using OpenTelemetry.
• Trace-Driven Analysis: Debugging production-like failures within Grafana.
• Efficiency & Scale: Managing the trade-offs of cardinality, sampling, and diagnostic value.

‍