Scholarlink

Modern software deployment increasingly relies on Continuous Integration and Continuous Deployment (CI/CD) pipelines, but maintaining consistent reliability and performance across these automated workflows remains a pressing challenge. This paper introduces a smart monitoring framework that combines Generative AI (GenAI) with Machine Learning (ML) to deliver proactive anomaly detection, intelligent alerting, and automated issue resolution. Unlike conventional systems that depend on rule-based alerts and post-failure diagnostics, our framework uses predictive modeling and a GenAI-powered recommendation engine (Ollama) to interpret pipeline behavior, forecast potential issues, and offer actionable solutions in real time. In parallel, a Ping API-based uptime monitoring service continuously evaluates service availability, reducing the risk of undetected outages. Through a combination of supervised and unsupervised learning models trained on real deployment logs, the system improves accuracy in identifying faults, reduces false positives, and streamlines response efforts. Results show a 30% reduction in system failures, a 25% drop in downtime, and a 45% improvement in resolution time over traditional methods. By integrating these AI-driven capabilities, the framework supports more reliable, self-healing CI/CD pipelines—advancing the goal of intelligent, automated DevOps environments. By integrating advanced AI techniques into CI/CD monitoring workflows, this framework offers a proactive and scalable solution that supports continuous service availability and paves the way for self-healing, AI-assisted DevOps systems.

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

A Generative AI and Machine Learning-Based Framework for Proactive CI/CD Pipeline Monitoring and Fault Remediation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Kuldeep Vayadande,
  • Yogesh Bodhe,
  • Aditri Sivakumar,
  • Aditya Bhattacharya,
  • Aditya Karad,
  • Arnav Anand,
  • Orison Bachute

摘要

Modern software deployment increasingly relies on Continuous Integration and Continuous Deployment (CI/CD) pipelines, but maintaining consistent reliability and performance across these automated workflows remains a pressing challenge. This paper introduces a smart monitoring framework that combines Generative AI (GenAI) with Machine Learning (ML) to deliver proactive anomaly detection, intelligent alerting, and automated issue resolution. Unlike conventional systems that depend on rule-based alerts and post-failure diagnostics, our framework uses predictive modeling and a GenAI-powered recommendation engine (Ollama) to interpret pipeline behavior, forecast potential issues, and offer actionable solutions in real time. In parallel, a Ping API-based uptime monitoring service continuously evaluates service availability, reducing the risk of undetected outages. Through a combination of supervised and unsupervised learning models trained on real deployment logs, the system improves accuracy in identifying faults, reduces false positives, and streamlines response efforts. Results show a 30% reduction in system failures, a 25% drop in downtime, and a 45% improvement in resolution time over traditional methods. By integrating these AI-driven capabilities, the framework supports more reliable, self-healing CI/CD pipelines—advancing the goal of intelligent, automated DevOps environments. By integrating advanced AI techniques into CI/CD monitoring workflows, this framework offers a proactive and scalable solution that supports continuous service availability and paves the way for self-healing, AI-assisted DevOps systems.