LF Observability study plan

Define the three pillars (metrics, logs, traces) and write a one-paragraph contrast of what question each answers

Classify ten example signals from a sample incident into the correct pillar and justify each choice

Diagram the path a request takes through a service and mark where each pillar would capture data

Explain cardinality, aggregation, and sampling trade-offs in your own words with two concrete examples each

Compare the four metric types (counter, gauge, histogram, summary) and choose the right type for five described scenarios

Compute rate, increase, and percentile estimates from a small hand-built sample dataset

Distinguish structured from unstructured logs and rewrite three unstructured log lines as structured JSON

Describe trace anatomy (span, parent-child, context propagation) and sketch a trace for a two-service call

Identify the three signals emitted by a single request and explain how to correlate them via trace ID and exemplars

Design log levels and retention rules for a service with high-volume debug output

Explain head-based versus tail-based sampling and state when each is appropriate

Diagnose three broken traces (missing parent, dropped context, clock skew) and propose fixes

Describe the OpenTelemetry architecture: API, SDK, exporter, collector, and receiver/processor/exporter pipelines

Instrument a small local app with the OTel SDK and emit at least one metric, one log, and one span

Configure a collector pipeline that receives OTLP, batches, and exports to a local backend

Explain resource attributes, semantic conventions, and context propagation across process boundaries

Add a collector processor that filters or redacts sensitive attributes and verify the effect in exported data

Compare automatic and manual instrumentation and apply manual spans to a function that auto-instrumentation misses

Tune batching and memory-limiter settings and record the effect on export volume

Propagate context across an HTTP boundary and confirm parent-child linkage in the backend

Compare pull versus push collection models and match each to a suitable deployment scenario

Define SLI, SLO, and error budget and derive an SLO for a described service from its metrics

Draft three alert rules that fire on symptoms rather than causes and justify each threshold

Explain dashboard design principles and critique a poorly designed dashboard against them

Trace a simulated incident from alert to root cause using metrics, then logs, then traces, and document each step

Calculate telemetry cost drivers and propose two concrete reductions without losing coverage

Describe the operational lifecycle: deploy, observe, alert, remediate, and postmortem

Explain how observability maturity differs from monitoring coverage and give two indicators of each

Re-derive the weakest three concepts from weeks 1-7 and write short explanations from memory

Complete a timed mixed-domain practice set and score it by domain

Re-run the week 4-5 demo end to end without notes to confirm retention

Build a final one-page cheat sheet covering pillar trade-offs, OTel pipeline stages, and SLO/alerting rules