NGINX Cookbook Advanced Recipes for High-performance Load Balancing (Derek Dejonghe) (Z-Library)

【One-Line Pitch】 A practical recipe collection for engineers who already understand HTTP, TCP, and modern multi-tier or microservice architectures and now need to configure NGINX or NGINX Plus for real production traffic. It is best suited to operators, SREs, and backend engineers who want proven patterns for load balancing, caching, security, cloud deployment, and observability rather than a from-scratch introduction to web serving. 【Book Arc】 - **Opening (~0%–9%)**: Establishes the cookbook format and the scope of both NGINX Open Source and NGINX Plus, then previews the full recipe map: load balancing, traffic management, caching, security, cloud deployment, App Protect, HTTP/3, OpenTelemetry, and njs. - **Early (~16%–34%)**: Covers the operational basics—process model, key files and commands, `include` organization, serving static content—then moves into high-performance load balancing with HTTP, TCP, and UDP upstreams, balancing methods, sticky cookies, and passive health checks. - **Middle (~38%–47%)**: Shifts to traffic management and scale: GeoIP/GeoIP2 access control, rate limiting, and massively scalable content caching with cache zones, cache locking, stale cache, and cache slicing for large media. - **Late**: The excerpts indicate later chapters on high-availability deployment modes, advanced activity monitoring, debugging and troubleshooting with access/error logs and request tracing, and performance tuning, but the sampled material does not cover their detailed recipes. - **Ending**: The book closes with performance tuning topics such as load-driver testing, browser and upstream connection reuse, response and log buffering, and OS-level tuning, followed by an index. 【Key Takeaways】 - **NGINX is configured as a hierarchy of contexts** (Early): `http`, `server`, and `location` blocks define what is listened for and how requests are routed; understanding this structure is the prerequisite for every later recipe. - **`include` statements are the main tool for keeping configuration maintainable** (Early): shared FastCGI or SSL snippets can be written once and reused across virtual servers, avoiding duplication in large configs. - **Load balancing is upstream-centric** (Early): the `upstream` module defines pools of Unix sockets, IPs, or hostnames, and per-server parameters such as weight control routing; methods include round-robin, least connections, least time, generic hash, random, and IP hash. - **Sticky sessions and health checks have open-source versus Plus trade-offs** (Early): sticky cookie affinity is an NGINX Plus feature, while passive health checking with `max_fails` and `fail_timeout` works in Open Source across HTTP, TCP, and UDP. - **Traffic management is policy-driven** (Middle): GeoIP/GeoIP2 variables can log, route, or restrict by country, and `limit_req_zone` with `limit_req` enforces request-rate limits, typically returning a 4xx status such as 429. - **Caching needs concurrency and freshness controls** (Middle): `proxy_cache_path` defines storage, `proxy_cache_lock` prevents duplicate cache fills, `proxy_cache_use_stale` serves expired content when upstreams fail, and cache slicing efficiently handles large byte-range resources like HTML5 video. - **Monitoring and debugging are first-class recipes** (Late): stub status, the NGINX Plus dashboard and API, OpenTelemetry, and Prometheus exporter cover metrics, while access logs, error logs, syslog forwarding, config debugging, and request tracing cover diagnosis. - **Performance tuning spans client, upstream, logging, and OS layers** (Ending): connection keepalive, response buffering, access-log buffering, and OS tuning are presented as complementary levers rather than a single fix. 【Reading Tips】 - Treat the early basics as a reference, not a tutorial: if you already run NGINX, skim the process-model and key-files recipes and deep-read the upstream, sticky, and health-check sections. - Read the load-balancing and caching chapters together, because upstream behavior and cache behavior interact heavily in production traffic patterns. - Pay attention to Open Source versus NGINX Plus labels; several recipes, including sticky cookie and some monitoring features, are Plus-only and will not work on the free build. - Use the later monitoring, debugging, and performance-tuning recipes as an operational checklist when you move a configuration from staging to production. - Keep the book nearby as a lookup cookbook: the value is in the specific directives and trade-offs, not in linear reading. 【Coverage Limits】 This guide is based on stratified excerpts that cover the opening, early, and middle material in detail, while the later chapters on high availability, monitoring, debugging, and performance tuning are only visible through the table of contents. Specific recipes, configurations, and examples from those later chapters are not covered here.