Formerly called 'Code to 35'. Add our main WeChat ID to access a wealth of shared resources (algorithms, interview prep, tech stacks: Java, Python, Go, big data). We mainly share serious development techniques, focusing on output-driven input. Occasionally we post life snippets and gossip. Our aim is to attract precise traffic and test advertising opportunities.
A production outage at 3 am revealed a 10‑second Full GC pause on an 8 GB heap; the article walks through evidence collection, heap‑size rebalancing, switching to G1, fine‑tuning G1 parameters, and code changes that together reduced Full GC pauses to 200 ms and improved latency dramatically.
The article explains how ZGC achieves sub‑millisecond pause times by embedding GC state in pointer bits (colored pointers) and using JIT‑injected load barriers that fix references on first access, comparing this approach with G1 and Shenandoah and detailing ZGC 2.0 enhancements and known pitfalls.
The article demystifies why Java's synchronized keyword can be slower by tracing the full path from CPU cache‑coherence (MESI) through bus and cache locks, volatile, biased and lightweight locks, up to the heavyweight monitor implementation, revealing each layer’s impact on visibility, ordering and atomicity.
At billion‑level request rates, the traditional cache‑aside pattern fails, so the article proposes a three‑layer architecture—Canal captures MySQL binlog, an MQ (RocketMQ/Kafka) buffers changes, and a coordinated local cache service updates L1 caches with short TTL, achieving high performance and eventual consistency.
The article explains why local locks fail in multi‑machine deployments, defines the three essential properties of a correct distributed lock, walks through seven Redis‑based lock solutions with code samples, highlights common production pitfalls, and provides a decision tree for selecting the right approach.
When a monitoring alarm showed the /api/order/create endpoint’s P99 latency jump from under 50 ms to over 3 s, the team traced the root cause to a Full GC storm caused by an unbounded static HashMap cache, using Arthas, heap dumps, MAT analysis, and finally applied both quick‑fix and permanent code changes to restore performance.
This guide walks you through the entire lifecycle of a RocketMQ message, identifies where loss can occur, provides step‑by‑step diagnostics using TrackType and timestamps, and offers concrete configuration, command‑line and recovery strategies for producers, brokers and consumers.
Java 21, the latest LTS release, adds fifteen JEPs but only five—virtual threads, record patterns, switch pattern matching, string templates, and sequenced collections—significantly alter coding style, offering lightweight concurrency, concise data deconstruction, modern switch syntax, powerful string interpolation, and enhanced collection APIs.
The article shows that interviewers don’t just want you to list design patterns; they assess your ability to identify real scenarios, solve concrete problems, justify trade‑offs, and discuss implementation details, illustrated with payment and order‑processing examples in Java and Spring.
The article argues that many senior developers only repeat the same one‑year‑old skills, showing that true experience requires deeper problem‑solving, trade‑off thinking, and the ability to eliminate issues rather than merely execute them.
