Lesson 01
Observability for Backend Systems Explained
Decision: map user impact to trustworthy metrics, traces, logs, and change evidence.
Production Observability & SRE topic cluster
Production Observability & SRE
A production path for trustworthy telemetry, Spring instrumentation, signals, JVM profiling, delivery pipelines, SLO alerting, on-call response, and verified recovery.
12 reading lessons ยท 1 capstone guide
Connect telemetry quality to verified user recovery.
Every lesson keeps actual user impact, observed evidence, alert state, mitigation, and verified recovery as separate conclusions.
-
-
Lesson 02
Telemetry Quality, Semantic Conventions, and Cardinality Explained
Define stable signal identity, units, timestamps, coverage, and bounded dimensions before trusting analysis.
-
Lesson 03
Spring Boot, Micrometer, and OpenTelemetry Instrumentation
Instrument the order service while preserving clear framework, agent, protocol, and Collector boundaries.
-
Lesson 04
Structured Logging and Correlation IDs Explained
Decision: preserve searchable event identity without leaking sensitive or unbounded data.
-
Lesson 05
RED, USE, and Golden Signals Explained
Decision: connect user symptoms to service behavior and resource saturation with bounded metrics.
-
Lesson 06
Distributed Tracing and Context Propagation Explained
Decision: propagate interoperable causal context while preserving sampling and trust boundaries.
-
Lesson 07
JVM Profiling and JFR for Production Backends
Use Java Flight Recorder as bounded profiling evidence for CPU, allocation, GC, locks, and virtual threads.
-
Lesson 08
OpenTelemetry Collector Pipeline Explained
Decision: operate a bounded, observable, vendor-neutral telemetry processing pipeline.
-
Lesson 09
SLI, SLO, Error Budgets, and Burn-Rate Alerts Explained
Decision: page on sustained user-impact and explicit error-budget consumption.
-
Lesson 10
Production Alerting, On-Call, and Runbooks
Turn actionable user symptoms into bounded pages, escalation, handoffs, and evidence-led runbooks.
-
Lesson 11
Incident Command, Communication, and Postmortems
Separate command, operations, and communications while preserving a factual timeline and owned follow-up work.
-
Lesson 12
Production Incident Troubleshooting with Logs, Metrics, and Traces
Decision: move from alert to tested hypothesis, reversible mitigation, and verified recovery.
-
Capstone lab
SLO Incident Triage Lab Examples
Compare actual and observed modeled impact, alert timing, mitigation windows, and the evidence needed before recovery can be claimed.