Observability Engineer learning path: everything to study, topic by topic

Core metrics collection platforms including Prometheus pull model, Datadog agent-based monitoring, and CloudWatch native AWS integration; metric types (counter, gauge, histogram, summary), push vs pull architectures, and metric naming conventions

Designing application-specific metrics that capture business and system health signals; RED (Rate, Errors, Duration) and USE (Utilization, Saturation, Errors) methodologies, metric granularity, instrumentation libraries, and avoiding metric explosion

Building effective dashboards in Grafana and similar tools; layout principles, variable templating, alert integration, dashboard-as-code with JSON/Grafonnet/Terraform, and audience-specific views (on-call, executive, team)

Managing high-cardinality metrics that threaten system stability and cost; label explosion detection, cardinality reduction strategies, relabeling pipelines, and long-term storage down-sampling with Thanos, Cortex, or Mimir

Implementing structured logging with JSON or key-value formats; log levels and severity, correlation IDs for request tracing, context enrichment, PII redaction, and logging standards across polyglot services

Centralizing logs with Elasticsearch-Logstash-Kibana (ELK), Grafana Loki, or CloudWatch Logs; log shipping agents (Fluentd, Fluent Bit, Filebeat), index management, log queries, and multi-tenant logging architectures

Creating alerts from log patterns and anomalies; log-derived metrics, pattern matching rules, threshold-based vs anomaly-based log alerts, false positive reduction, and integrating log alerts with incident management workflows

Managing log storage costs through retention policies, tiered storage, log sampling, compliance-driven retention requirements, cold storage archival, and cost-per-GB optimization across logging platforms

Instrumenting applications with OpenTelemetry SDKs for traces, metrics, and logs; auto-instrumentation vs manual instrumentation, context propagation (W3C Trace Context, B3), exporters, and collector architecture

Using distributed traces to diagnose latency, errors, and bottlenecks; flame graphs, waterfall views, span attribute filtering, trace-to-log correlation, and root cause analysis using trace data

Automatically discovering and visualizing service dependencies from trace data; dependency graph generation, critical path identification, blast radius analysis, and detecting hidden dependencies and circular calls

Controlling trace volume through sampling without losing visibility into errors and tail latency; head-based vs tail-based sampling, probabilistic sampling, rate-limiting, priority sampling, and the OpenTelemetry Collector sampling processor

Designing high-signal, low-noise alerts; symptom-based vs cause-based alerting, alert severity classification, deduplication, grouping, inhibition rules, alert fatigue metrics, and sustainable on-call practices

Defining SLOs and using error budget burn rate alerts; multi-window burn rate alerting, SLI selection, error budget tracking dashboards, and the relationship between SLOs, error budgets, and release decisions

Encoding incident response procedures as executable runbooks; runbook design, automated remediation workflows, self-healing systems, integration with alerting pipelines, and measuring time-to-remediation improvements

Correlating metrics, logs, and traces to accelerate root cause analysis; exemplar linking, trace-to-metric correlation, log-to-trace bridging, unified observability views, and building correlation workflows in Grafana/Datadog

Architecting data pipelines for telemetry collection, transformation, and routing; OpenTelemetry Collector pipelines, vector/fluentd routing, data transformation and enrichment, fan-out to multiple backends, and pipeline reliability

Using eBPF for low-overhead kernel-level observability; network tracing, syscall monitoring, application profiling without instrumentation, tools (Cilium, Pixie, bpftrace), and eBPF-based service mesh observability

Capturing real user experience metrics from browsers and mobile apps; Core Web Vitals, session replay, error tracking, user journey mapping, synthetic vs real user monitoring, and connecting frontend signals to backend traces

Driving observability adoption across engineering teams; developer on-boarding to observability tools, building an observability guild, defining standards and golden signals, measuring adoption maturity, and making observability a first-class engineering practice

Optimizing the total cost of observability infrastructure; data volume forecasting, tiered ingestion pricing, metric and log volume reduction, vendor cost comparison, and building chargeback models for observability spend

Evaluating observability vendors and planning migrations; RFP/RFI processes, proof of concept design, data portability with OpenTelemetry, migration strategies, multi-vendor architectures, and avoiding vendor lock-in