{"categories":[{"label":"Monitoring","url":"https://skillfed.io/packages/category/system-monitoring/4"}],"enrichment":{"capability":"Provides core telemetry abstractions (spans, events, attributes) and a recorder interface for collecting and exporting observability data from applications and long-running jobs.","skillfed_tags":["telemetry","distributed-tracing","observability"],"use_cases":["Instrument ML training jobs to track checkpointing, model evaluation, and other domain-specific events with automatic timing","Collect distributed traces from microservices by creating spans for operations and exporting them to a collector pipeline","Add telemetry to long-running batch jobs with configurable exporters (Kafka, Weights & Biases, OTEL collector) without tight coupling","Build domain-specific observability libraries on top of the core abstractions for specialized application types","Aggregate and filter telemetry events at the recorder level (e.g., computing checkpoint statistics across multiple events)"],"what_it_does":"nv-one-logger-core is a telemetry collection library that provides the foundational abstractions for instrumenting applications\u2014particularly long-running jobs like ML training. It defines core concepts (spans representing units of work, events representing singular points in time, and attributes as properties) modeled on OpenTelemetry standards. The library intentionally keeps its core lightweight by minimizing third-party dependencies, allowing it to be adopted without conflicting with existing application dependencies.\n\nThe package provides a Recorder interface that applications use to start and stop spans, record events, and report errors. It also includes a flexible exporter configuration system supporting direct code configuration, YAML/JSON files, and Python entry points, with clear priority rules. Applications can use it via higher-level domain-specific libraries, the Recorder interface directly, a context manager (timed_span), or by instantiating exporters manually\u2014though the Recorder approach is recommended to reduce instrumentation mistakes.","worth_installing":"Yes, if you need a lightweight, vendor-agnostic telemetry foundation for applications or jobs. The core is intentionally minimal to avoid dependency conflicts. However, the aging maintenance status (289 days since last release) suggests limited active development\u2014verify whether it fits your stability and support expectations before committing to it as a long-term instrumentation backbone."},"id":"nv-one-logger-core","links":{"html":"https://skillfed.io/packages/nv-one-logger-core","md":"https://skillfed.io/packages/nv-one-logger-core.md","pypi":"https://pypi.org/project/nv-one-logger-core/"},"maintenance":{"status":"aging"},"meta":{"latest_release":"2025-10-29","license_spdx":null,"license_treatment":"permissive","name":"nv-one-logger-core","python_support":"supports_current","summary":"Extensions to onelogger library to use Open telemetry (OTEL) as a backend."},"popularity":{"monthly_downloads":231034,"position":9099,"tier":"top_15000"},"security":{"n_vulnerabilities":0},"version":"2.3.1"}