Blog

Mert Özen's blog posts on software development, backend, frontend, and DevOps.

Series

.NET Core Web API 36 Angular 31 JavaScript Basics 19 JavaScript Intermediate 18 JavaScript Advanced 16 TypeScript Basics 14 TypeScript Intermediate 14 TypeScript Advanced 12 Entity Framework Core 28 Dapper 16 Git & GitHub 29 Web Security (OWASP Top 10) 21 LLM Fundamentals 20 C# Fundamentals 20 Object-Oriented Programming in C# 16 C# Collections and Generics 11 LINQ 17 Asynchronous Programming in C# 17 Advanced C# 14 C# Versions 6 Solution Patterns 22 SOLID and Design Principles 12 Clean Code 13 Design Patterns 23 UML and Modeling 18 Introduction to Software Architecture 9 Clean, Onion and Hexagonal Architecture 8 Vertical Slice Architecture 6 Domain-Driven Design 12 CQRS and Event Sourcing 10 Messaging and Event-Driven Architecture 9 Modular Monolith 2 Microservices 10 UI Architecture Patterns 9 Architecture Decisions 7 System Design 15 Distributed Systems Fundamentals 11 Background Jobs and Scheduled Jobs 11 Working with Message Queues 14 Logging in .NET 10 Caching 10 Real-time Applications 9 ASP.NET Core MVC and Razor Pages 13 CI/CD Fundamentals 8 GitHub Actions 16 Jenkins 12 Web Servers and Hosting 12 Deployment Strategies 9 Kubernetes 19 Authentication and Authorization 18 ASP.NET Core Security 8 Frontend Security 9 API Security (OWASP API Top 10) 11 AI with .NET 14 RAG 13 AI Agents and MCP 14 Test Strategy 4 Unit Testing 13 SDLC (Software Development Life Cycle) 6 Agile Mindset 6 Scrum 11 Kanban 7 Introduction to Product Management 9 SEO Fundamentals 7 Technical SEO 12 GEO (Generative Engine Optimization) 7 Algorithms and Data Structures 22 Debugging and Problem Solving 13

2 posts found #performance

Newest Oldest Popular

Backend PART 01

Caching Strategy: What to Cache, at Which Layer and for How Long

A caching decision starts with measurement: in Bookstore the category menu, with 2 queries per request and a 19.9 ms p95, was the only real candidate, and HybridCache cut p95 to 3.2 ms and 4,000 queries to 0, while the cheap, per-user cart should not be cached. TTL is derived from freshness requirements, and each layer's invalidation owner is decided up front.

8 Nov 2026 10 min 6

Backend PART 02

IMemoryCache in Depth: SizeLimit, Eviction, Change Tokens and the Hidden Stampede in GetOrCreate

In an IMemoryCache with a full SizeLimit, an entry that exceeds the limit is never cached: book 101 added to a 100-unit cache was rejected, and background compaction evicted Low priority entries first, then the least recently used, bringing the cache down to 85. GetOrCreateAsync isn't atomic; 100 concurrent requests on an empty cache ran the factory 100 times, while striped SemaphoreSlim locks and HybridCache cut that to 1. Expiration has no timer; group invalidation uses change tokens.

8 Nov 2026 15 min 4