Search

Lessons (40)

• Train Your First Model from Scratch

  … slope x + intercept) print(f"slope={slope:.3f}, intercept={intercept:.3f}") print(f"baseline test mse={baseline error …

• Capacity Planning, SLOs and Error Budgets

  … For request-based SLOs, calculate the allowed failed fraction from eligible requests instead. 3. Load-test peaks and …

• Linked Lists

  … Fast & slow pointers (tortoise and hare) Two pointers move through the list: slow moves 1 step, fast moves …

• Two Pointers & Sliding Window

  … remove duplicates in place (fast & slow) js // [1, 1, 2, 2, 2, 3] → first 3 slots become [1 …

• Resilience Patterns

  … Without it, one slow service causes a cascading failure : mermaid flowchart LR U[👥 Users] -- G[Gateway] -- O[Orders …

• Code Profiling

  Find the code that is actually slow or retaining memory before you optimise. Learn a repeatable profiling workflow …

• Debug Latency Like an Investigator

  … Compare healthy and slow traces; locate the span that gained time. 3. Inspect queue age, pool acquisition, query …

• Observability: Logs, Metrics & Traces

  … Learn the three pillars, the golden signals, SLOs and OpenTelemetry. The big idea A car dashboard shows your …

• Next.js Rendering: CSR, SSR, SSG, ISR & Server Components

  … JSON Note over B: ✅ page visible (late) ✅ Rich interactivity, cheap hosting (static files). ❌ Slow first paint on slow …

• 10,000 Users Do Not Need 10,000 Database Connections

  … an average query time can hide a small number of very slow queries that occupy the pool for …

• Backend Design Scenarios

  … One downstream service becomes slow and suddenly the whole application is slow. How do you stop one dependency …

• Resilient Service Calls: Retries, Breakers and Load Shedding

  Keep one slow dependency from turning a small outage into a full-system outage. The problem in context …

• CloudWatch: Know What Your App Is Doing

  … Search affected request IDs and trace slow/error paths. 4. Use the approved mitigation: rollback or restore capacity …

• L6 Production Debugging Casebook

  … Compare one slow distributed trace with a healthy trace for the same route. Find the span that gained …

• Releases, Observability and Production Safety

  … CI validates an immutable artifact, controlled rollout observes SLOs, and a rollback path limits blast radius](/img/system …

• Design a Logging and Monitoring Platform

  … Alerts should use user-facing SLO symptoms and link to runbooks. Preserve tenant isolation, retention controls and an …

• 500 Users Upload Documents at Once: Design the Background Job Pipeline

  Move slow document processing out of the HTTP request with object storage, a job table, a task queue …

• Storage Engines & Database Operations

  … Bound pool size, reserve capacity for critical traffic, set deadlines and investigate long transactions or slow queries. Backups …

• Timeouts, Retries & Backpressure

  A slow dependency should consume a bounded budget, not create an infinite retry storm. The problem in context …

• Observability and Incident Response

  … views of the same request Imagine checkout becomes slow at 2:05 PM. - Metrics say p99 latency rose …

• Indexes, EXPLAIN & Query Plans

  … writes become slow and memory pressure rises. - Putting a low-selectivity flag first: most rows still match. - Wrapping …

• Database Indexing & Transactions

  … and / < fast. Creating and checking indexes (PostgreSQL) sql -- Slow: no index on email SELECT FROM users WHERE email …

• Frontend Performance & Core Web Vitals

  … Real users on slow phones matter most. Fixing LCP: show the main content sooner mermaid flowchart LR T …

• React Internals & Concurrent Rendering

  … Measure and fix the expensive path if it remains slow. Suspense and streaming Suspense shows a fallback while …

• Monolith vs Microservices

  … one table breaks 3 services", "one slow service makes everything slow". Conway's Law 📘 "Organisations design systems that …

• Proxies, APIs and Real-Time Protocols

  … messages. How to prove it worked Test disconnects, slow clients and rolling restarts. Measure reconnect success, update lag …

• Arrays & Strings

  … See how arrays live in memory, why some operations are instant and others slow, and the essential string …

• Caching & Redis

  … flush a non-production cache and confirm the application becomes slower temporarily, not incorrect or unavailable. Two related …

• Views, Triggers & Stored Procedures

  … but row-by-row cursor updates are often slow. Prefer set-based updates, clear ownership and auditability. Use …

• Design a URL Shortener

  … able to see clicks without making every redirect slow. Example requirements: 10,000 redirect requests per second at …

• What Is Clean Code?

  … a little slower at the start, much faster later. This hidden cost is called technical debt . Like a …

• YAGNI: You Aren't Gonna Need It

  … Error handling and security Caching layer before measuring slowness Code that is easy to change later 💡 The secret …

• Concurrency, Resilience & Distributed Workflows

  … when only one run is allowed. Backpressure means slowing producers or shedding optional work when consumers cannot keep …

• AWS Messaging: Queues, Topics and Events

  … A queue lets checkout finish after the order is safely accepted, while a worker handles the slow task …

• Go Web Services with Gin

  … 5 time.Second, // protects against slow-loris clients ReadTimeout: 10 time.Second, WriteTimeout: 15 time.Second, IdleTimeout: 60 …

• Clean Architecture in Depth

  … I/O, external services, time, randomness Every CRUD screen goes through 6 layers Slow development Keep simple reads …

• Backend Languages: Node.js, Go, Python & Rust

  … CPU threads) AI/ML, data, scripting, fast prototyping Slower raw speed Rust async (Tokio) + threads, no data races …

• Docker & Containers

  … handling sacks, barrels and crates of every shape: slow and chaotic. The standard shipping container changed everything: any …

• DSA Interview Patterns Cheat Sheet

  … 3. Brute force: say it out loud with its complexity, even if it's slow. It shows you …

• Consumers & Consumer Groups

  … be sorted. One person can do it alone, slowly. A team can split the pile so each person …