Best of OpenTelemetrySeptember 2024

  1. 1
    Article
    Avatar of hnHacker News·2y

    Open-source DataDog + PostHog for AI agents / RAG apps. Fast, reliable and insightful. Written in Rust 🦀. YC S24.

    Laminar is an OpenTelemetry-based framework designed for easy instrumentation of LLM and vector DB calls. Featuring semantic events-based analytics, Laminar turns pipeline outputs into insightful metrics. Built with Rust and incorporating RabbitMQ, Postgres, and Clickhouse, it aims to provide scalable and efficient event and trace analytics. Users can start quickly with a managed platform or set up locally using Docker Compose.

  2. 2
    Article
    Avatar of last9Last9·2y

    OpenTelemetry Collector: The Complete Guide

    OpenTelemetry provides a vendor-agnostic observability framework suitable for distributed systems. The OpenTelemetry Collector, or otelcol, allows for the receipt, processing, and export of telemetry data in diverse formats. It supports multiple protocols and can replace the need for multiple monitoring agents. Key benefits include vendor neutrality, data transformation, scalability, and cost reduction. Use cases range from microservices monitoring to legacy system integration. Advanced topics cover custom processors, high availability setups, and performance tuning. The guide also compares the Collector with Jaeger and emphasizes best practices for effective deployment.

  3. 3
    Article
    Avatar of prometheusPrometheus·2y

    Prometheus 3.0 Beta Released

    Prometheus 3.0-beta is now available for testing, featuring a completely rewritten UI, enhancements to Remote Write 2.0, expanded OpenTelemetry support, and experimental Native Histograms. Users are encouraged to test the beta and report any issues for a more stable final release. Notable additions include support for UTF-8 characters in metric and label names, and new configurations for OTLP ingestion.

  4. 4
    Article
    Avatar of devopsDevOps.com·2y

    OpenTelemetry Isn’t the Hero We Need: Here’s Why it’s Failing our Stack

    OpenTelemetry was intended to be a unified standard for collecting and correlating traces, logs, and metrics from distributed systems. However, it has become bloated and inefficient due to corporate influence and feature creep. In contrast, eBPF, operating at the kernel level, provides real-time, low-overhead observability and granular insights without the same drawbacks. While OpenTelemetry offers a big-picture view useful for distributed tracing, eBPF excels in detailed system performance insights. Combining both could provide comprehensive observability.

  5. 5
    Article
    Avatar of last9Last9·2y

    OpenTelemetry Filelog Receiver: Collecting Logs from Kubernetes

    The OpenTelemetry filelog receiver is a valuable tool for log collection in Kubernetes environments. It reads log files and converts them into the OpenTelemetry log format. Setting it up involves deploying the OpenTelemetry Collector in the Kubernetes cluster, configuring it via a YAML file, and applying the configuration. Optimization tips include using batch processors and memory limiters, and troubleshooting advice covers common issues like missing logs and parsing errors. The receiver integrates well with the OpenTelemetry ecosystem and can handle various log sources, including syslog. Advanced parsing techniques and integration with backends like Last9 Levitate are also discussed.