The knight that protects your servers. An Observability and AIOps platform designed to detect anomalies and predict infrastructure failures before users even notice them.
Modern applications generate massive volumes of logs and operational events. When something breaks, engineering teams often spend hours investigating logs and metrics to identify the root cause.
Igris is an observability and AIOps platform designed to proactively detect anomalies in system behavior by analyzing real-time log streams using Machine Learning models.
Instead of reacting to failures, Igris predicts them.
The platform continuously monitors infrastructure signals and alerts engineers when abnormal patterns appear — often before a failure impacts users.
High-throughput ingestion pipeline designed to process large volumes of system and application logs.
Machine Learning models such as Isolation Forests and Autoencoders detect subtle behavioral anomalies in infrastructure data.
Receive alerts via push notifications, dashboards, or webhooks before failures escalate.
A centralized interface to visualize logs, anomalies, alerts, and infrastructure health.
Igris follows an event-driven architecture where logs are ingested, processed, analyzed by ML models, and then transformed into actionable insights.
Servers / Applications
│
▼
Log Ingestion API (Fastify)
│
▼
Event Stream Processing
│
▼
ML Engine (Python)
│
▼
Anomaly Detection
│
▼
Alerts + Dashboard
The platform separates responsibilities into independent services to ensure scalability and flexibility.
This project is structured as a monorepo to facilitate collaboration, code sharing, and scalability across services.
igris-platform
│
├── apps
│ ├── api # Log ingestion service (Node.js + Fastify)
│ ├── web # Observability dashboard (React)
│ └── mobile # Mobile alerting app (React Native)
│
├── services
│ └── ml-engine # Machine learning worker (Python)
│
├── packages
│ ├── core # Shared business logic
│ ├── types # Global TypeScript types
│ ├── database # Database schema and ORM configuration
│ └── configs # Shared linting and tooling configs
│
└── docs
├── adr # Architectural Decision Records
└── architecture
Shared packages allow all services to reuse types, schemas, and core logic while maintaining service independence.
- Node.js
- Fastify
- TypeScript
- React
- React Native
- Python
- Isolation Forest
- Autoencoders
- Drizzle ORM
- Event-driven architecture
- Microservices
- Monorepo with shared packages
Igris is currently in early MVP development.
Initial work is focused on building the core infrastructure:
- Log ingestion pipeline
- Event processing architecture
- Machine learning anomaly detection
- Observability dashboard
Documentation and setup instructions will evolve as the project matures.
- Log ingestion API
- Basic anomaly detection
- Initial dashboard
- Alert system
- Advanced anomaly detection models
- Infrastructure metrics integration
- Alert routing
- Historical anomaly analysis
- Predictive infrastructure analytics
- Automated incident detection
- AI-assisted root cause analysis
- Automated remediation workflows
The long-term vision of Igris is to evolve into a fully autonomous AIOps platform capable of:
- Predicting infrastructure failures before they happen
- Automatically detecting incidents
- Assisting engineers in root cause analysis
- Providing intelligent remediation suggestions
Ultimately, Igris aims to reduce operational complexity and empower teams to focus on building reliable systems.
Setup instructions will be added as the MVP progresses.
- Node.js (v18+)
- Python (3.9+)
- Package manager (pnpm, npm, or yarn)
Contributions are welcome!
As the project evolves, a detailed CONTRIBUTING.md guide will be provided.
For now:
- Open an issue describing your idea
- Discuss architecture or improvements
- Submit a pull request
This project is licensed under the Apache 2.0 License.
See the LICENSE file for more information.