0 views collected around this technical thread.
This article introduces the OpenAI Agents SDK, explains its key capabilities such as Agent Loop, Handoffs, Guardrails, and Tracing, provides practical Python code examples, compares it with other multi‑agent frameworks, and discusses best practices for building reliable AI applications.
This article explains why observability is essential for Spring AI applications, outlines common cost‑control and performance challenges, and provides a step‑by‑step guide—including Maven setup, client configuration, service implementation, metric exposure, Zipkin tracing, and architecture insights—to create a fully observable, enterprise‑grade AI translation service.
This article introduces SigNoz, a self‑hosted open‑source observability platform that unifies metrics, logs, and traces, outlines its core capabilities, shows how to install it with Docker, and compares its resource efficiency to commercial solutions like DataDog and Elastic.
OpenTelemetry unifies tracing and metrics by merging OpenTracing and OpenCensus, offering vendor‑neutral APIs, SDKs, and a collector that standardize telemetry data collection, context propagation, and export to various back‑ends, with detailed components such as Tracer, Meter, and shared Context layers.
This article explains how Spring Cloud automatically wraps managed thread pool beans with trace-enabled proxies to preserve distributed tracing information, details the ExecutorBeanPostProcessor implementation, shows the relevant configuration and instrumentation code, and notes that manually created executors must be wrapped manually.
This article provides a comprehensive guide to Linux's ftrace tool, explaining its purpose, various tracers, how to set up and use it via debugfs, detailed command examples, implementation details, practical use cases for performance tuning and debugging, and a comparison with other tracing utilities.
This article explains the built‑in debugging and tracing methods of the rule engine, including the debug API, trace operations, stack‑management functions such as caller checks, stack formatting, and thread‑stack tracing, along with usage examples and special cases for controlling output.
The article introduces the Cool Request IDEA plugin, explains its tracing capabilities for arbitrary packages, automatic MyBatis function monitoring, customizable timing colors, script-based environment manipulation, and provides a Java code example for handling responses, highlighting its usefulness for backend developers.
This article explains how to replace invasive manual timing code with a lightweight Java Agent that uses the java.lang.instrument API, ASM bytecode manipulation, and Attach/Arthas tools to automatically measure method execution time and perform dynamic class transformation at runtime.
Observability turns a black‑box application into a system by gathering logs, metrics, and traces, using alerts to spot anomalies, then linking trace IDs to logs; OpenTelemetry standardizes this with instrumented client agents, a Collector (receivers, processors, exporters), and backend storage, while Java agents, span propagation, exemplars, eBPF, and bundles like SigNoz or OpenObserve let teams choose between a custom OTel stack or a solution.
The article outlines practical backend interface performance optimizations—including proper indexing, SQL tuning, parallel remote calls, batch queries, asynchronous processing, scoped transactions, fine-grained locking, pagination batching, multi-level caching, sharding, and monitoring tools—to dramatically reduce latency and improve throughput.
This article details Huya's transition from fragmented, non‑standard monitoring solutions to a unified OpenTelemetry‑based platform, covering project background, pain points, design decisions, SDK architecture, data pipeline, storage, alerting, root‑cause analysis, and future plans, highlighting the benefits of standardization and zero‑cost service integration.
Observability, defined as measuring system state through logs, metrics, and tracing, enhances stability of distributed architectures by enabling rapid fault detection, deeper insight, and proactive issue resolution, distinguishing it from traditional monitoring and supporting DevOps, SRE, and business objectives.
This article explains how containerization and Kubernetes gave rise to micro‑services, why observability (metrics, logging, tracing, profiling) became essential for managing their complexity, and how OpenTelemetry’s unified data model enables seamless integration of these signals across modern cloud‑native stacks.
Perfetto’s tracing system links multiple producers to a single consumer via shared‑memory buffers, where careful sizing of pages, chunks, and central buffers, along with tuned protobuf encoding and scheduling priorities, mitigates CPU overhead and prevents data loss, enabling reliable observability on Android devices.
Trace2.0 is an OpenTelemetry‑based application monitoring system that processes petabyte‑scale trace data using multi‑channel client protocols, gRPC, load‑balancing optimizations, ZSTD compression, Kafka pipelines, ClickHouse storage, and a JDK 21 upgrade with virtual threads, achieving significant performance and cost improvements.
Observability in distributed systems combines event logging, aggregated metrics, and request tracing—each offering distinct trade‑offs in detail, storage, and overhead—and while the ELK stack dominates log and metric handling, tracing solutions such as EagleEye and SkyWalking differ by protocol and language, prompting many teams to adopt unified, cloud‑native platforms like Alibaba Cloud’s Log Service for lower cost, real‑time analysis and simplified management.
This article details the systematic investigation of intermittent availability drops in a high‑throughput Go service, covering hypothesis formulation, extensive profiling with pprof, gctrace, strace, fgprof, go trace, heap analysis, the discovery of a gcache LFU bug, and the final remediation steps.
This article explores why logs are essential in software development, outlines when to record them, discusses the value of logging in large-scale distributed systems, and examines the capabilities required of log‑operation tools such as APM, metrics, tracing, ELK, Prometheus, and custom batch querying solutions.
The investigation revealed that the Go http.Client was not reusing connections because response bodies were closed before being fully read, causing the underlying connections to be marked dead; fully reading the bodies restored pooling, eliminated DNS‑resolution timeouts and dramatically lowered query‑per‑second spikes.