In continuous integration / continuous delivery (CI/CD) work environments, integration and deployment of software is a daily process. Automated tests are conducted on each integration attempt in controlled environments, generating logs that document the outcomes, indicating whether tests are passed or failed. However, software engineers need to manually analyse the generated logs, which is a time-consuming task, requiring significant effort from the engineers. This research introduces the CI Nurse Agent, a system designed to automate the analysis of logs by developing an agentic Retrieval-Augmented Generation(RAG) system. The system uses rule-based retrieval methods, vector-based retrieval mechanisms and a Open AI gpt-4o to diagnose failures. The experimental results show that the CI Nurse Agent achieves high accuracy in diagnosing the failures with a success rate of 91.67% in root cause identifications. These findings conclude that the system can reduce the need for manual analysis and offer an alternative to manual log analysis in CI/CD workflows.