Cloud Data Catalog Metrics via Event-Driven Aggregation

The patent inverts the traditional pull-based metrics collection approach by implementing a push-based approach where resource events are automatically published to a message queue and workers consume events to update metrics. This reversal eliminates the need for continuous polling queries while maintaining architectural simplicity through event-driven architecture.

The patent introduces a message queue as an intermediary component between resource event sources and metrics processing workers. This mediator decouples the production of resource events from their consumption and processing, enabling scalable metrics collection without direct point-to-point queries between components.

The patent implements a self-service mechanism where resource event sources automatically publish status change events to the message queue when their state changes. This eliminates the need for external polling systems to continuously query resource status, reducing query overhead while maintaining metrics freshness through event-driven updates.

The patent ensures continuous metrics updates through an event-driven architecture where resource status changes trigger immediate event publication and processing. This continuous useful action replaces periodic polling, maintaining up-to-date metrics without generating excessive queries by only processing when changes occur.

The patent segments the message queue processing by introducing partitioned queues and multiple workers that process different segments of events in parallel. This segmentation enables complete metrics coverage across all partitions while reducing overall processing time through concurrent event processing across multiple worker instances.

The patent implements selective event processing where workers process subsets of events from the message queue based on filtering criteria and priority levels. This partial action approach processes only the most critical or relevant events first, achieving sufficient metrics coverage without processing every single event, thereby reducing total processing time while maintaining reliability for key metrics.