This series finale presents a universal 6-step SOP for Java production troubleshooting, a bottom-up layered diagnosis model, a quick-reference guide for 10 common failure patterns with symptoms, root causes, and tools, plus five golden principles for incident handling.
This guide covers real-world diagnosis and resolution of Java thread deadlocks and thread pool exhaustion — the two most common causes of service hangs with zero errors — using jstack, Arthas, and thread-state analysis to pinpoint root causes and apply permanent fixes.
The article explains how JDK Flight Recorder (JFR) provides low‑overhead, continuous JVM event recording that outperforms traditional snapshot tools like top, jstack, and jmap, and offers practical commands and use‑cases for diagnosing CPU spikes, GC issues, thread contention, and latency problems in production Java services.
A production incident where core API latency spiked from 50ms to 30s+ due to connection leaks and exhausted MySQL max_connections is dissected, revealing how default HikariCP settings become time bombs under load and how to build a defense system with leak detection, parameter tuning, slow SQL interception, and kill mechanisms.
This article explains why traditional monitoring tools often fail in production, introduces Arthas as a lightweight, non‑intrusive Java diagnostic solution, and walks through five real‑world scenarios—slow interfaces, thread blockage, memory leaks, hot‑fixes, and data inconsistency—showing exact commands, code snippets, and visualizations to quickly locate and resolve root causes.
The author documents three StarRocks production issues: varchar(50) truncating MaxCompute strings, BE memory spikes caused by API gateway timeout mismatches, and a global dictionary error fixed by disabling low-cardinality optimization.
This article introduces Arthas, an open‑source Java diagnostic tool, and demonstrates how to install it, use its command‑line and web console, and apply common commands such as dashboard, thread, watch, trace, and tt to quickly locate CPU spikes, deadlocks, memory leaks, and other production problems without redeploying code.
This article compiles a comprehensive list of common production problems encountered with Apache Flink, covering cluster sizing, checkpoint failures, backpressure analysis, resource allocation, deployment errors, UDF definitions, data skew, Kafka configurations, and provides detailed troubleshooting steps and best‑practice recommendations.
This guide demonstrates how to replace traditional debugging with Arthas, covering installation of IDEA plugins, quick 5‑minute problem resolution, hot‑deployment using ArthasHotSwap, and advanced time‑tunnel commands for recording and replaying method executions.
