Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Application-level observability for adaptive Edge to Cloud continuum systems

Kaddour Sidi, Daniel Balouek, Baptiste Jonglez

TL;DR

The paper addresses the challenge of maintaining performance objectives in heterogeneous Edge-to-Cloud (E2C) systems by proposing an application-level observability framework that pairs developer-driven instrumentation with an SLO-aware feedback loop. The approach combines OpenTelemetry, Prometheus, K3s, and Chaos Mesh to enable fine-grained monitoring and autonomous adaptation at both infrastructure and application levels, demonstrated through a video-processing use case. Key contributions include an end-to-end observability pipeline with metric scalability, a two-stage feedback mechanism (Status and Causes Inference; Reconfiguration and Resolution), and the ability to add or redefine metrics at runtime. Preliminary results show improved scalability, fault tolerance, and responsiveness, with application-specific metrics such as frame rate and appearance rates guiding adaptive actions to preserve latency, throughput, and accuracy in dynamic workloads. Overall, the work provides a practical pathway toward resilient, self-optimizing E2C continuum applications that can operate with reduced human intervention in real-world deployments.

Abstract

Modern Edge-to-Cloud (E2C) systems require fine-grained observability to ensure adaptive behavior and compliance with performance objectives across heterogeneous and dynamic environments. This work introduces an application-level observability framework that integrates developer-driven instrumentation and SLO-aware feedback for autonomous adaptation. By combining OpenTelemetry, Prometheus, K3s, and Chaos Mesh, the framework enables real-time monitoring and adaptive control across the continuum. A video processing use case demonstrates how application-level metrics guide automatic adjustments to maintain target frame rate, latency, and detection accuracy under variable workloads and injected faults. Preliminary results highlight improved scalability, fault tolerance, and responsiveness, providing a practical foundation for adaptive, SLO-compliant E2C applications.

Application-level observability for adaptive Edge to Cloud continuum systems

TL;DR

The paper addresses the challenge of maintaining performance objectives in heterogeneous Edge-to-Cloud (E2C) systems by proposing an application-level observability framework that pairs developer-driven instrumentation with an SLO-aware feedback loop. The approach combines OpenTelemetry, Prometheus, K3s, and Chaos Mesh to enable fine-grained monitoring and autonomous adaptation at both infrastructure and application levels, demonstrated through a video-processing use case. Key contributions include an end-to-end observability pipeline with metric scalability, a two-stage feedback mechanism (Status and Causes Inference; Reconfiguration and Resolution), and the ability to add or redefine metrics at runtime. Preliminary results show improved scalability, fault tolerance, and responsiveness, with application-specific metrics such as frame rate and appearance rates guiding adaptive actions to preserve latency, throughput, and accuracy in dynamic workloads. Overall, the work provides a practical pathway toward resilient, self-optimizing E2C continuum applications that can operate with reduced human intervention in real-world deployments.

Abstract

Modern Edge-to-Cloud (E2C) systems require fine-grained observability to ensure adaptive behavior and compliance with performance objectives across heterogeneous and dynamic environments. This work introduces an application-level observability framework that integrates developer-driven instrumentation and SLO-aware feedback for autonomous adaptation. By combining OpenTelemetry, Prometheus, K3s, and Chaos Mesh, the framework enables real-time monitoring and adaptive control across the continuum. A video processing use case demonstrates how application-level metrics guide automatic adjustments to maintain target frame rate, latency, and detection accuracy under variable workloads and injected faults. Preliminary results highlight improved scalability, fault tolerance, and responsiveness, providing a practical foundation for adaptive, SLO-compliant E2C applications.
Paper Structure (23 sections, 3 equations, 7 figures, 2 tables)

This paper contains 23 sections, 3 equations, 7 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Background and Related Work
  3. Observability Driven Feedback Mechanism
  4. Status and Causes Inference
  5. Metrics Collection
  6. Metrics Preprocessing
  7. Critical Metrics Extraction
  8. Status Inference
  9. Reconfiguration and Resolution
  10. Action Inference
  11. Reconfiguration
  12. Evaluation
  13. Knowledge Database
  14. Novelty in Application-Specific Metrics and Developer-Driven Instrumentation
  15. Evaluation
  16. ...and 8 more sections

Figures (7)

  • Figure 1: Feedback Mechanism Scheme
  • Figure 2: Overview of the Edge-to-Cloud Use-Case
  • Figure 3: Observability Workflow for the Use-Case Application on Grid5́000
  • Figure 4: System evolution over time showing processing time, metrics, and pod count; the scale-out window indicates adaptive resource adjustment under load.
  • Figure 5: Detected motions vs response time (AR: 05, 05, 05)
  • ...and 2 more figures