Skip to main content

Courses About Pricing Educators

Sign in Get started

Courses About Pricing Educators

Get started Sign in

Loading courses

Loading lessons

Resilience and Observability: System Design: Uncharted

Courses System DesignResilience and Observability

Resilience and ObservabilityPro

Budget reliability, bound dependency failures, see request paths, and learn from incidents.

System Design
Unit 10 of 12
20 lessons
149 min
420 XP

This unit is part of Pro. Upgrade to open every lesson in it.

Lessons

20 total

1Everything FailsAdopt failure as a design assumption and bound each dependency failure so the system can recover without cascading collapse.ProInteractiveConcept7 min20 XPOpen
2SLOs, Not VibesWrite a measurable service-level objective over a time window and relate its allowed misses to an error budget.ProInteractiveConcept7 min20 XPOpen
3Napkin: The NinesCalculate and combine yearly downtime budgets for multiple availability targets using the stated per-target rounding rule.ProInteractiveConcept7 min20 XPOpen
4p50 Lies, p99 TellsInterpret latency percentiles to find a user-visible tail violation that a healthy median can hide.ProInteractiveConcept8 min20 XPOpen
5The Latency BudgetAdd serial hop latencies and quantify how far an end-to-end request exceeds its latency budget.ProInteractiveConcept7 min20 XPOpen
6The Missing TimeoutDiagnose worker exhaustion caused by a hung dependency and identify the finite timeout that bounds the wait.ProInteractiveConcept7 min20 XPOpen
7Retry StormsUse exponential backoff and jitter to reduce retry frequency and break synchronization during dependency recovery.ProInteractiveConcept7 min20 XPOpen
8When Not to RetryDecide when retries are unsafe or futile based on idempotency, error permanence, overload, and a bounded retry budget.ProInteractiveConcept7 min20 XPOpen
9The Circuit BreakerExplain how closed, open, and half-open circuit-breaker states stop futile calls and probe for recovery safely.ProInteractiveConcept7 min20 XPOpen
10BulkheadsIsolate critical and optional dependencies in separate bounded resource pools so one slow path cannot exhaust the other.ProInteractiveConcept7 min20 XPOpen
11Degrade GracefullyApply a preplanned degradation ladder that sheds increasingly optional work while preserving the core user action.ProInteractiveConcept7 min20 XPOpen
12Shed at the DoorExplain why rejecting excess demand at the gateway protects admitted requests and spends fewer scarce downstream resources.ProInteractiveConcept7 min20 XPOpen
13Metrics, Logs, TracesChoose metrics, logs, or traces according to whether an incident question concerns aggregate behavior, one event, or one cross-service request.ProInteractiveConcept7 min20 XPOpen
14One Request, Many ServicesReconstruct a request across services using a shared trace identifier and parent-child span relationships.ProInteractiveConcept7 min20 XPOpen
15Alert on PainChoose urgent alerts from user-visible objective violations rather than resource signals that may be healthy under load.ProInteractiveConcept7 min20 XPOpen
16Find the Bad DeployCorrelate a deployment marker with the first sustained post-deploy latency-budget violation.ProInteractiveConcept8 min20 XPOpen
17The PostmortemStructure a blameless postmortem from user impact and timeline through contributing conditions, recovery lessons, and assigned actions.ProInteractiveConcept7 min20 XPOpen
18Harden the BakeryDesign a checkout path that balances requests, decouples notification work with a queue, and provides a fallback notification handler.ProDiagramProject10 min20 XPOpen
19Explain the StormExplain the retry-storm feedback loop and how timeouts, jittered backoff, circuit breaking, and retry budgets limit it.ProInteractiveConcept8 min20 XPOpen
20Resilience Mastery CheckEvaluate incident controls for bounded waiting, desynchronized retries, resource isolation, and user-centered alerting.ProInteractiveMastery check10 min40 XPOpen

Storage and Media Blueprints