SentinelOps

Inspiration

Modern software development moves fast, but broken CI/CD pipelines slow down entire teams. Developers often spend hours sifting through thousands of lines of raw runner logs to pinpoint simple build failures, syntax errors, or dependency mismatches. I built SentinelOps to automate pipeline troubleshooting directly inside the developer's workspace using GitLab Duo Chat.

What it does

SentinelOps is an autonomous AI agent integrated into GitLab Duo Chat that monitors, diagnoses, and remediates CI/CD pipeline failures in real time.

• Automated Log Analysis: Intercepts failed runner jobs (such as docker_build failures) and extracts key execution errors.
• Intelligent Root Cause Diagnosis: Analyzes raw logs alongside codebase context to identify dependency conflicts, missing scripts, or configuration flaws.
• Actionable Remediation: Generates instant, ready-to-use fixes (e.g., corrected Dockerfiles, updated dependencies, and syntax adjustments) right within Duo Chat.
• Green DevSecOps: By accelerating error resolution, SentinelOps prevents redundant failed build iterations, reducing overall compute waste and energy consumption.

How I built it

• Backend Microservice: Built using Node.js, Express.js, and TypeScript, structured around clean microservices patterns.
• GitLab CI/CD Integration: Engineered a full 5-stage DevSecOps pipeline (test -> secure -> package -> deploy -> monitor).
• AI & Diagnostic Engine: Leveraged GitLab Duo Chat and automated log processing to parse runner outputs and suggest intelligent fixes.
• Deployment & Monitoring: Deployed automatically to Google Cloud Run with post-deployment health verification.

Challenges I faced

• Log Parsing Precision: Isolating true root-cause error signals from verbose runner outputs required careful log structuring and context extraction.
• Seamless Duo Integration: Ensuring the agent provided actionable, copy-pasteable Dockerfiles and configurations directly inside GitLab Duo Chat without unnecessary noise.

What I learned

Building SentinelOps highlighted the power of embedding AI directly into developer workflows rather than relying on external dashboards. Integrating diagnostic AI at the CI/CD layer dramatically shortens feedback loops and enhances productivity.

What's next for SentinelOps

• Automated pull/merge request generation for proposed pipeline fixes.
• Predictive failure warnings prior to running full DevSecOps stages.

Built With

• docker
• express.js
• gitlab-ci
• gitlab-duo
• google-cloud-run
• mongodb
• node.js
• react
• tailwindcss
• typescript