Avvikit

Avvikit is a software toolkit designed for deviation analysis and anomaly detection in time-series data and process monitoring. The name is derived from the Norwegian word avvik, meaning deviation, reflecting its focus on identifying departures from expected behavior. Avvikit is intended for use in industrial analytics, manufacturing quality control, energy monitoring, and related domains.

The toolkit provides data ingestion from multiple sources, including CSV files, JSON APIs, databases, and real-time streams. It offers data preprocessing features such as normalization, resampling, and feature extraction, along with a range of anomaly detection methods. These include statistical approaches like z-score and robust scaling, seasonal-trend decomposition, and machine-learning-based models. Avvikit also includes a rule engine to express custom deviation criteria, and it supports alerting through email, webhooks, or integration with incident management systems. A lightweight visualization dashboard helps users inspect detections, and a modular plugin architecture enables easy extension. It exposes a REST API for integration with external systems and supports common storage backends and message brokers.

Architecture-wise, Avvikit comprises a core engine that coordinates modules, adapters for various data sources, reusable algorithm plugins, the rule engine, and a user interface. It is designed to be pluggable, with components for storage, caching, and data visualization decoupled to support flexible deployments. Typical deployments include on-premises or cloud-based setups, connected to dashboards such as Grafana or custom UIs.

Licensing and availability: Avvikit is released as open-source software under a permissive license and is available on public code-hosting platforms, accompanied by documentation and community-driven support.