CloudSentinel AI

• 

  The dashboard that shows the investigation stats cards, severity heatmaps and quick actions

• 

  Investigation page - paste messy logs to the AI engine and it extracts signals, finds correlations, and builds a structured incident report.

• 

  Investigation history page - view list of completed investigations

• 

  AI assistant - go deeper and get more insightful reports or answers about your logs

Inspiration

Cloud infrastructure failures can be difficult and time-consuming to debug. Developers often spend hours searching through logs trying to understand what actually caused an issue. We wanted to build a tool that makes incident investigation faster, clearer, and more intelligent using AI.

What it does

CloudSentinel AI is an AI-powered infrastructure incident investigation platform. Users can paste or upload cloud and system logs from sources like Kubernetes, Docker, CI/CD pipelines, and backend services. The platform analyzes the logs, detects anomalies, identifies likely root causes, classifies severity, and recommends possible fixes. It also generates incident summaries, investigation timelines, and infrastructure insights to help developers troubleshoot issues more efficiently.

How we built it

We built CloudSentinel AI as a fullstack web application using modern frontend and backend technologies. The frontend provides an observability-style dashboard for viewing investigations, AI findings, severity levels, and incident timelines. The backend handles:

• log ingestion
• anomaly detection
• signal extraction
• incident history
• AI-powered analysis workflows

We integrated Mistral AI for:

• root cause analysis
• infrastructure explanations
• incident summaries
• troubleshooting guidance

Challenges we ran into

One of the biggest challenges was making the AI responses technically accurate and useful instead of generic. We also had to distinguish between: actual root causes, downstream symptoms and repeated infrastructure failures. Another challenge was designing a workflow that felt like a real investigation platform rather than a simple log viewer.

Accomplishments that we're proud of

We are proud of building a platform that can turn unstructured infrastructure logs into structured and understandable incident investigations.

We are especially proud of the AI-generated root cause analysis, correlated failure detection, investigation timelines and actionable infrastructure insights

What we learned

We learned a lot about infrastructure observability, incident investigation workflows, AI-assisted debugging and designing developer-focused tools. We also learned how important context and correlation are when troubleshooting cloud systems.

What's next for CloudSentinel AI

Next, we want to expand CloudSentinel AI with real-time log streaming, cloud provider integrations, collaborative investigations, Slack/Discord alerts, infrastructure topology mapping and AI-powered remediation suggestions. Our goal is to evolve CloudSentinel AI into a complete AI-native observability and incident response platform.

Built With

• deno
• gemini
• react
• shadcn/ui
• supabase
• tailwindcss
• typescript