Technical Breakdown
ObservoAI is an advanced observability and API watchdog platform built for the Barclays Hack-o-Hire 2025. It monitors distributed systems across on-prem, cloud, and multi-cloud environments, utilizing a 3-tier machine learning pipeline for anomaly detection. Re-imagining operations, it features natural language log/metric chatting, Gemini-powered root cause analysis (RCA), and automated voice-call alerts during high-priority incidents.
Engineering Logic
The system processes high-frequency telemetry data asynchronously. The 3-tier ML models run in separate backend threads to maintain real-time evaluation with minimal latency, triggering notifications through an autonomous dialer agent.
Aesthetic Precision
We implemented a dark-themed observability dashboard with glowing, high-contrast status nodes and trace cards, allowing developers to spot path anomalies and critical error states immediately.
Technical Features & Superpowers
3-Tier Anomaly Detection: Employs ARIMA for trend forecasting, LSTM for deep degradation tracking, and Isolation Forest for black-swan events.
AI Voice-Call Alerts: Automatically triggers phone calls to on-call engineers via AI voice agents when critical thresholds are crossed.
Query Logs via Chat: Ask questions about logs and metrics using natural language, converted dynamically to PromQL and TraceQL.
Root Cause Analysis (RCA): Utilizes Gemini to generate deep incident diagnostics and sends automated reports via email.
High-Frequency Bank Simulator: Simulates heavy banking transaction journeys to validate collection thresholds and stress-test pipelines.