Technology
Multi-Tenant Architecture in ERP Systems: The Anatomy of Sharing
My experiences and strategic decisions while designing a multi-tenant architecture for a manufacturing ERP. Sharing models, data isolation, and performance…
#performance
48 posts found.
Technology
High Cardinality Metrics: Does the Benefit Outweigh the Cost?
Examining the impact of high cardinality metrics on system performance, cost analysis, and optimal usage scenarios.
Technology
Database Indexes: Necessary for Every Query?
I examine when database indexes are beneficial, when they hurt performance, and the right indexing strategies with real-world scenarios.
Career
Cardinality Explosion: Should Every Detail Really Be Observed? And
What is cardinality explosion in monitoring systems, why does it happen, and how does this situation affect both systems and an engineer's career? Practical...
Career
Why Does Using an ORM Decrease Database Performance? An Experience...
I explain how the convenience of ORMs negatively affects database performance, especially in enterprise applications, using my own field experiences.
Technology
Metric Cardinality: An Overlooked Performance Burden or a Developer
How does metric cardinality affect system performance? In this guide, we delve deep into overlooked burdens and developer mistakes.
Technology
RED Metrics Design: Service-Oriented or Workflow-Oriented?
Should RED metrics be designed based on services or workflows? This post explores the pros, cons, and best use cases for each approach.
Career
AI Agent Tool-Use: Boundaries in Cost and Performance Balance
I provide a pragmatic perspective by examining the cost and performance limits of AI agents' tool usage with real-world scenarios.
Career
Database Partitioning Cost: Is It Really Worth It?
I analyze the benefits and costs of database partitioning. When should you partition, and when should you avoid it? I share my experiences.
Technology
ORM Tools Are Overrated: Why They Fall Short in Large-Scale Projects?
I examine the shortcomings of ORM tools in large-scale projects, their performance bottlenecks, and alternative approaches with concrete examples.
Technology
LLM Inference Caching: How to Balance Cost and Latency?
I explain the intricacies of LLM inference caching and what to consider when balancing cost and latency, with practical examples.
Tutorials
Anatomy of Database Index Structures: Fundamentals of Query
A detailed examination of database index structures (B-tree, GIN, BRIN) and strategies for enhancing query performance. With real-world scenarios and concrete.
Life
✍️ Hand-written
Swap Fire on My VPS: A Nightmare That Started with a Kernel CVE Patch
I detail the process that began with my VPS's swap usage suddenly spiking and the system crashing, including the kernel CVE patch and the steps I took to.
Tutorials
✍️ Hand-written
Swap Fire: My Kubernetes Experiment on a 7.6 GB VPS
A pragmatic analysis of swap memory issues and their solutions encountered while experimenting with Kubernetes on a small VPS.
Tutorials
Why Are My Docker Containers Slow? A Monitoring Guide for My Own VPS
A practical guide to monitoring the performance of Docker containers on your own VPS and finding the root causes of slowdowns. Systemd, cgroup, and journald…
Tutorials
Guide to Detecting and Limiting Resource-Hog Containers on a VPS
I'm sharing a step-by-step guide on how I identified resource consumption issues on my own VPS and applied limits to Docker containers.
Technology
SQLite and Concurrency: The Lockout Experienced at islistesi.com
A first-hand account of the SQLite concurrency and lockout problems I faced in the islistesi.com project, with the solution steps and lessons learned.
Technology
Cloudflare Cache's Blind Spot: The Cost of Bypass Rules
I explain the unexpected effects of Cloudflare cache bypass rules and how I overcame them with Nginx to improve performance. My experiences on my own VPS.
Technology
Kernel Memory Wars: The Hidden Swap Trap and Its Solutions
Want to understand the hidden swap trap on Linux systems and learn memory management strategies for high-performance systems? Detailed…
Technology
Stealth Resource Contention in Containers: Problems and Solutions
Learn about stealth resource contention issues in containerized environments and effective solutions to this complex problem.
Career
The Curse of Stale Cache in High-Traffic Applications: Strategies and…
Learn how stale data hurts performance in high-traffic applications and the ways to break out from under that curse.
Technology
Database Connection Leaks in Production: The Quiet Resource Wars
Connection leaks in production are a sneaky threat — they drain system resources without anyone noticing and quietly tank performance. In this post we look at…
Technology
Service Mesh Sidecar Overhead: A Hidden Performance Tax
I dig into the hidden performance costs of the service mesh sidecar pattern — resource consumption, latency, and operational cost — and how to reason about…
Technology
Cold Start in Serverless Apps: A Hidden Performance Trap
I take a deep dive into the Cold Start problem in serverless architectures — why it happens, what it does to performance, and how to actually dodge it…
Tutorials
The Fragility of the Distributed Database Shard Key
I unpack the critical role of the shard key in distributed databases, the risks it carries (hotspots, data skew), and the strategies to keep that fragility…
Tutorials
The Anatomy of Unscalable Database Decisions in System Architecture
A deep look at the long-term effects of database choices in system architecture and the scalability traps they create. The cost of bad decisions and…
Technology
The Virtual Network Gateway Performance Mystery: A Hidden…
We investigate the overlooked performance bottlenecks of virtual network gateways in production. This article covers why they matter, the hidden problems…
Tutorials
Redis Sharding: The Hidden Wars in Production and Its Dark Side
Explore the complexity, challenges, and hidden production battles of Redis sharding. We shed light on the dark side of sharding.
Technology
✍️ Hand-written
Cloudflare HTML Cache Stuck at 1.1%: Recovery with Nginx map
Cloudflare cache was stuck at 1.1%. Astro Node adapter returns max-age=0 for HTML. Override based on content-type via nginx map directive.
Technology
The Silent Betrayal of Reverse Proxy Buffer Settings
Discover the hidden impact of reverse proxy buffer settings on performance and security. Optimization tips and tricks on the Mustafa Erbay blog!
Career
The Zombie Process Hunt in Production: Anatomy of a Hidden…
A detailed look at the 'zombie process' problem in production environments and how to analyze and resolve this hidden form of resource waste.
Tutorials
The Cache Invalidation Dead End in Large-Scale Systems
An in-depth look at cache invalidation problems frequently encountered in large-scale systems and the solutions that actually work.
Technology
Virtual NIC Queues: The Hidden Performance Killer
Learn how virtual network interface queues hurt network performance and how I get past this hidden bottleneck.
Tutorials
A Hidden Resource Exhaustion War: The Deadly Dance of Containers
Learn about the hidden resource-exhaustion war containers fight, and how to manage this deadly dance. Performance optimization and stability included…
Career
Hidden Performance Issues in the Shadow of Service Mesh: For Your…
Beyond the advantages Service Mesh offers, the often-overlooked performance costs and how they reflect on a software engineer's career…
Technology
The 'Thundering Herd' Problem in Distributed Systems: Anatomy of a…
Take a deep look at the 'Thundering Herd' problem that threatens performance and stability in distributed systems. Understand this destructive effect and…
Career
Packet Loss in a Multi-Layer Network: Fighting a Performance Killer
Learn the causes of packet loss in multi-layer networks and how to deal with this hidden performance killer. Optimize your network performance.
Technology
Distributed Tracing Issues in Critical Systems: The Anatomy of…
Take a deep dive on Mustafa Erbay's blog into the complexity of distributed tracing in critical systems and the invisible errors that come with it…
Technology
Database Provisioning Mistakes in the Cloud and How to Fix Them
A deep look at database provisioning mistakes I keep running into on cloud platforms, the symptoms they cause, and the fixes that actually hold up in…
Technology
Retry Storms: Timeout Budget and Latency Amplification
In distributed systems, badly designed retries make outages worse. An approach to limiting damage with timeout budgets, retry budgets, and backpressure.
Tutorials
Cache Stampede in Front of the CDN: Origin Server Loading Wars
Explore the Cache Stampede problem in front of CDNs, its causes, and effective strategies to avoid overloading the origin server.
Tutorials
Linux SoftIRQ Saturation and IRQ Affinity Runbook
Quick triage, measurement and safe tuning steps (ring, queue, IRQ, RPS) under packet drops, high softirq load and ksoftirqd pressure.
Tutorials
SLO-Driven Load Testing with k6: Capacity Baselines and Release Gates
A practical approach that turns load testing from a peak-RPS race into an SLO-driven (latency/error) capacity baseline and a CI release gate.
Technology
Cache Stampede (Thundering Herd) and Operational Defenses
A guide to taming the stampede (thundering herd) risk that can crush a backend after TTL expiry or a cache flush — using jitter, singleflight, and stale…
Technology
SLO-Based Degrade Modes and Load Shedding
Producing controlled loss instead of a random collapse when a system is under pressure: rate limits, queues, feature flags and prioritization.
Technology
Database Connection Pool Saturation and the Latency Feedback Loop
A practical framework to detect the queue, timeout, and retry loop that emerges when a connection pool clogs, and to intervene safely.
Tutorials
Linux Conntrack Capacity Planning and Alerting Runbook
A practical guide for generating signals before the nf_conntrack table fills up, applying safe sysctl tuning, and recovering in a controlled way during an…
Tutorials
Linux TCP Backlog and SYN Flood Resilience Runbook
A runbook to triage the connect timeout crisis when the SYN backlog/accept queue fills up, apply rapid mitigation, and design lasting resilience.