altiquum-event-engine is a production-grade, highly concurrent, and durable event-driven processing engine written in Go. It establishes the message routing, event ingestion, and reliable delivery backbone for the ALTIQUUM ecosystem.
altiquum-event-engine/
├── go.mod # Go module declaration (links to altiquum-core)
├── README.md # Technical specifications
├── LICENSE # MIT License
├── Makefile # Test and benchmark execution
├── event/ # Event schemas, structures, and validation
├── dedupe/ # Idempotency and deduplication registry
├── store/ # Binary segment Write-Ahead Log (WAL)
├── broker/ # Ingesting, header-filtering, routing registry
├── delivery/ # Concurrency workers, retry-backoff, and DLQ
├── telemetry/ # Telemetry metrics and trace spans
├── benchmarks/ # Core broker and ingestion benchmarks
└── examples/ # End-to-end crash-recovery simulation
- Scalable, non-blocking ingestion channel using Go channels and worker routing pools.
- Concurrent subscription matching based on event headers.
- Events are appended to a binary Write-Ahead Log (WAL) before dispatch.
- If a consumer fails or is interrupted mid-delivery, the broker can replay the WAL segment on restart.
- Committing events (acknowledgments) mark WAL positions.
- Subscribers can filter events based on headers (e.g. key-value matching), allowing flexible pub/sub patterns without complex topics structures.
- Integrated idempotency filters tracking processed event IDs within a sliding time-to-live (TTL) window to block duplicate deliveries.
- Delivery pipelines feature automated exponential backoff with full jitter for temporary failures.
- Unprocessable events that exhaust all retries are permanently routed to a Dead-Letter Queue (DLQ) directory for off-line audit.
- Go 1.26+
altiquum-coreinitialized in/scratch/altiquum-core(linked locally viago.modreplacement)
Run tests:
make testRun benchmarks:
make benchRun simulated workload:
go run examples/main.go