Eventdiffer

Eventdiffer is a software concept and library designed to diff two sequences of events to identify additions, deletions, modifications, and reordering between streams. It is used in event-sourced architectures, data reconciliation, and debugging distributed systems where time-stamped events flow through multiple components. By modeling events as records with an identifier, a type, a timestamp, and an optional payload, eventdiffer provides a structured view of how histories diverge and where reconciliation is possible.

It supports multiple diff strategies, including exact matching by id, type, and payload, as well as windowed or fuzzy matching that aligns events within a configurable time window. It detects out-of-order deliveries, gaps, and duplicates, and can produce patches describing created, deleted, or modified events, as well as mappings between equivalent events across streams. The library offers batch and streaming APIs, allowing diffs on complete logs or incremental updates, and outputs a machine-readable diff representation such as JSON that highlights changes and suggested reconciliations.

Internally, eventdiffer comprises a lightweight event model, a core diff engine, and pluggable matching strategies plus adapters for common transport and storage backends. The engine can operate in memory for small datasets or in streaming mode with bounded memory through windowing and checkpointing. Extensions may include custom similarity metrics, patch formats, and integration with event stores or message queues.

Common use cases include reconciling two event stores after failures, validating delivery guarantees between microservices, auditing changes in an event log, and debugging complex event-driven workflows.

Although the term has appeared in discussions of event-centric data reconciliation, there is no single universal implementation. The concept relates to event sourcing, diff and patch algorithms, sequence alignment, and data reconciliation practices in distributed systems.