Scholarlink

<p>The growing demand for cross domain service deployments has intensified the need for robust telemetry solutions in distributed environments such as the cloud to edge continuum. This paper presents an extended and validated telemetry architecture capable of dynamically collecting, processing, and exporting metrics across heterogeneous domains. Built to integrate with existing distributed orchestration frameworks, the solution enables runtime telemetry reconfiguration through a policy/intent-based approach. A key innovation is the introduction of a lightweight telemetry agent that provides runtime adaptability by modifying telemetry pipelines without service interruption. The system supports both local and remote monitoring, ensuring observability even when services are offloaded to external infrastructure. Evaluation in real world use cases confirms the effectiveness and responsiveness of the solution, achieving configuration updates in as little as 3.18&#xa0;s. While developed within the FLUIDOS framework, the proposed telemetry service is flexible enough to be adapted to other distributed computing platforms.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Cross-domain telemetry architecture: dynamic observability in the computing continuum

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月9日

查看全文
  • José Manuel Bernabé Murcia,
  • Alejandro Molina Zarca,
  • Stefano Braghin,
  • Antonio Skármeta

摘要

The growing demand for cross domain service deployments has intensified the need for robust telemetry solutions in distributed environments such as the cloud to edge continuum. This paper presents an extended and validated telemetry architecture capable of dynamically collecting, processing, and exporting metrics across heterogeneous domains. Built to integrate with existing distributed orchestration frameworks, the solution enables runtime telemetry reconfiguration through a policy/intent-based approach. A key innovation is the introduction of a lightweight telemetry agent that provides runtime adaptability by modifying telemetry pipelines without service interruption. The system supports both local and remote monitoring, ensuring observability even when services are offloaded to external infrastructure. Evaluation in real world use cases confirms the effectiveness and responsiveness of the solution, achieving configuration updates in as little as 3.18 s. While developed within the FLUIDOS framework, the proposed telemetry service is flexible enough to be adapted to other distributed computing platforms.