Уровень 0 · материалов: 4
В кластер входят документы, описывающие изменение архитектуры хостинга и внутренние настройки параметров веб-сервера для устранения узких мест в производительности.
Общие признаки: переход с общего хостинга на VPS, изоляция процессов веб-сервера, настройка лимитов потоков и соединений, выбор серверного стека для повышения производительности
Группа выше: Развёртывание и настройка серверов
Смысл: The main idea is to demonstrate that by using a lightweight stack (Nginx + PHP-FPM) on a cloud VPS instead of traditional shared hosting, users can achieve superior performance, better security through isolation, and lower costs for MODx Revolution sites.
A technical tutorial on configuring a high-performance, low-cost cloud hosting environment for MODx Revolution using Ubuntu, Nginx, and PHP-FPM with user-based isolation.
Смысл: The text illustrates how moving from shared hosting to a VPS and implementing memory caching can drastically resolve performance bottlenecks for a LiveStreet CMS site, while highlighting the value of professional networking within tech communities.
The author shares how a connection on Habr led to migrating their slow LiveStreet site from shared hosting to an optimized VPS, resulting in a massive speed increase.
Смысл: The main idea is that isolating web server processes (providing dedicated Apache instances per user/site) is more effective for managing shared hosting loads than global resource limiting, provided there is sufficient physical RAM to support the increased memory overhead.
DiPHOST.Ru solved unpredictable traffic spikes and resource contention by moving from a shared Apache pool to dedicated Apache instances per site, supported by a hardware RAM upgrade.
Смысл: The main idea is that standard ASP.NET/IIS default settings are often insufficient for high-load applications that rely heavily on external web services, and performance can be significantly improved by manually tuning thread limits and connection quotas.
A technical guide on tuning IIS and ASP.NET thread and connection limits to prevent request queuing and bottlenecks in high-load environments.