Stage 4 · Provision
System Design for SRE
Real-world architecture decisions, reliability patterns, and distributed systems design for production at scale.
11 modules66 lessons7.7h of reading66/66 live today
Prerequisite
Applications + infrastructure.
What this course leaves you with
- Design systems end-to-end on a whiteboard
- Reason about reliability, latency, and trade-offs
- Model architectures that survive scale
System DesignSREDistributed SystemsArchitecture
Why this stage matters — You can design — now run those workloads on Kubernetes.
Progress
0 of 66 lessons complete. Progress is stored locally in your browser so you can pick the path back up later.
0%
course completion
01
6 lessons
Designing for High Availability
Redundancy, failover, and eliminating single points of failure.
- 01Availability FundamentalsSLAs, SLOs, SLIs — the language of reliability.6 min
- 02Redundancy PatternsActive-passive, active-active, and quorum-based designs.8 min
- 03Failover StrategiesAutomatic vs manual, DNS vs load balancer, stateful failover.7 min
- 04Multi-Region ArchitectureActive-active across regions, data replication, latency tradeoffs.8 min
- 05Chaos EngineeringHypothesis-driven failure injection to validate resilience.7 min
- 06Runbook AutomationTurning incident response into executable code.6 min
02
6 lessons
Load Balancing & Traffic Management
Distributing traffic intelligently for performance and resilience.
- 01Load Balancing AlgorithmsRound-robin, least connections, consistent hashing, and when to use each.7 min
- 02L4 vs L7 Load BalancingTCP/UDP vs HTTP/gRPC — tradeoffs and use cases.6 min
- 03Traffic Splitting & CanaryGradual rollouts, A/B testing, and feature flags at the infrastructure layer.7 min
- 04Service Mesh Traffic ControlIstio/Linkerd for mTLS, retries, timeouts, and circuit breaking.8 min
- 05API Gateway PatternsAuthentication, rate limiting, request transformation at the edge.6 min
- 06Global Load BalancingGeo-DNS, anycast, and latency-based routing across continents.7 min
03
6 lessons
Database Scaling Strategies
From vertical scaling to sharding — choosing the right approach.
- 01Vertical Scaling & Read ReplicasWhen to scale up vs out, synchronous vs async replication.6 min
- 02Sharding StrategiesHash, range, directory-based — and how to reshard without downtime.8 min
- 03Caching LayersRedis, Memcached, local caches — cache-aside, read-through, write-through.7 min
- 04Distributed SQL (NewSQL)CockroachDB, Spanner, TiDB — strong consistency at scale.7 min
- 05Polyglot PersistenceChoosing the right database for each access pattern.6 min
- 06Zero-Downtime MigrationsDual-write, CDC, and the strangler fig pattern for database migrations.8 min
04
6 lessons
Observability-Driven Design
Building systems that tell you when they're breaking.
- 01Logs, Metrics, TracesThe three pillars and why structured logging matters.5 min
- 02SLO-Driven DevelopmentDefining error budgets and making reliability a product feature.7 min
- 03Distributed TracingOpenTelemetry, context propagation, and debugging across services.7 min
- 04Alerting PhilosophySymptom-based vs cause-based, reducing alert fatigue.6 min
- 05Dashboard DesignRED and USE methods, golden signals, and operational dashboards.6 min
- 06Cost as a SignalTracking cloud spend per service, per request, per team.5 min
05
6 lessons
Data Systems at Scale
Designing data paths for throughput, durability, and predictable failure modes.
- 01Partitioning ModelsHash, range, and composite sharding with hot-key mitigation and online resharding.8 min
- 02Replication TopologiesLeader-follower, multi-leader, and quorum replication with failover tradeoffs.8 min
- 03Consistency ModelsLinearizability, read-your-writes, eventual consistency, and CAP/PACELC decisions.7 min
- 04Cache HierarchiesBrowser, CDN, edge, Redis, and application caches with invalidation strategies.7 min
- 05CDN & Edge DesignAnycast, origin shielding, stale-while-revalidate, and cache-key design.6 min
- 06Event Streams & QueuesKafka, RabbitMQ, SQS, and exactly-once semantics with idempotent consumers.8 min
06
6 lessons
System Design Interviews
Practicing crisp requirements, estimates, APIs, data models, and trade-off narratives.
- 01Back-of-the-Envelope EstimationQPS, storage, bandwidth, cache size, and fanout estimates using powers of ten.6 min
- 02Design a URL ShortenerBase62 keys, collision handling, redirects, analytics, and abuse prevention.7 min
- 03Design a Rate LimiterToken bucket, leaky bucket, sliding window logs, and Redis Lua enforcement.8 min
- 04Design a News FeedFanout-on-write, fanout-on-read, ranking, denormalized timelines, and cache warming.8 min
- 05Design a Chat SystemWebSockets, presence, ordering, offline delivery, and message storage partitions.8 min
- 06Trade-off FramingLatency versus consistency, cost versus durability, and explaining rollback plans.6 min
07
6 lessons
Asynchronous Systems
Designing queues and streams that absorb bursts without losing work.
- 01Queue SemanticsVisibility timeouts, acknowledgements, dead-letter queues, and retry policies in SQS and RabbitMQ.7 min
- 02Event Streaming with KafkaTopics, partitions, consumer groups, offsets, and compaction in Kafka clusters.8 min
- 03Idempotent ConsumersDeduplication keys, transactional outboxes, and exactly-once boundaries for event handlers.7 min
- 04Backpressure ControlRate limits, bounded buffers, consumer lag, and load shedding for asynchronous workers.7 min
- 05Workflow OrchestrationTemporal, Step Functions, sagas, and compensation steps for long-running business processes.8 min
- 06Async ObservabilityTracing message hops, queue depth, consumer lag, and poison-message handling with OpenTelemetry.6 min
08
6 lessons
Resilience Patterns
Containing failures with explicit budgets, isolation, and degraded modes.
- 01Timeout BudgetsConnection, request, and overall deadline budgets across Envoy, gRPC, and HTTP clients.7 min
- 02Retry StrategiesExponential backoff, jitter, retry budgets, and idempotency requirements for safe retries.7 min
- 03Circuit BreakersOpen, half-open, and closed states with failure thresholds in Hystrix and Envoy.6 min
- 04Bulkheads and IsolationThread pools, connection pools, resource quotas, and cell-based blast-radius reduction.7 min
- 05Graceful DegradationFallback data, partial responses, static modes, and brownout controls during dependency outages.6 min
- 06Failure Injection TestsValidating resilience with Toxiproxy, Chaos Mesh, latency injection, and dependency blackholes.8 min
09
6 lessons
Scalability & Capacity
Forecasting growth and proving systems can handle planned demand.
- 01Horizontal vs Vertical ScalingCPU, memory, disk IOPS, and network ceilings that determine scale-up or scale-out choices.6 min
- 02Stateless Service DesignExternal sessions, shared caches, sticky routing risks, and immutable application instances.6 min
- 03Capacity ModelingQPS, concurrency, Little's Law, saturation curves, and headroom targets for production services.8 min
- 04Load Testing MethodsOpen-loop and closed-loop tests with k6, Locust, wrk, and realistic traffic mixes.8 min
- 05Autoscaling SignalsCPU, queue depth, custom metrics, KEDA, and Kubernetes HPA stabilization windows.7 min
- 06Cost Performance TradeoffsRightsizing, reserved capacity, spot instances, and latency SLOs in cloud capacity plans.7 min
10
6 lessons
Cache Design Patterns
Using cache layers deliberately while preserving correctness under change.
- 01Cache-Aside PatternRead-through misses, write invalidation, TTLs, and stale reads with Redis and Memcached.7 min
- 02CDN Cache KeysVary headers, query normalization, surrogate keys, and origin shielding in Fastly and CloudFront.7 min
- 03Invalidation StrategiesPurge APIs, versioned objects, write-through updates, and event-driven invalidation pipelines.8 min
- 04Thundering Herd ControlRequest coalescing, probabilistic early refresh, locks, and stale-if-error responses.8 min
- 05Hot Key MitigationKey salting, local caches, replication, and load-aware routing for skewed access patterns.7 min
- 06Cache ObservabilityHit ratio, eviction rate, latency, object size, and origin offload metrics for cache tiers.6 min
11
6 lessons
Design Case Studies
Combining reliability tradeoffs into complete production architectures.
- 01Design a Notification SystemFanout queues, user preferences, provider retries, deduplication, and delivery receipts.8 min
- 02Design a Metrics PipelineScraping, remote_write, aggregation, cardinality limits, and long-term storage with Thanos.8 min
- 03Design a Search ServiceIndexing pipelines, Elasticsearch shards, freshness targets, and relevance ranking updates.7 min
- 04Design a File Storage ServiceMultipart uploads, object metadata, checksum validation, lifecycle policies, and CDN delivery.7 min
- 05Design a LeaderboardRedis sorted sets, partitioned rankings, periodic snapshots, and anti-cheat data checks.6 min
- 06Design an Audit LogAppend-only storage, tamper evidence, retention policies, and queryable compliance exports.7 min