Event model
Every observation in xrelia is represented as an event. Health checks, synthetic requests, metrics, dependency updates, changes, and incidents all follow a unified event model.
This allows xrelia to correlate signals across services and platforms, detect patterns, and build a complete operational timeline.
Unified event taxonomy
Section titled “Unified event taxonomy”All monitoring data is classified into a small set of event types.
The primary event categories are health check events, synthetic request events, metric events, dependency events, change events and incident events.
Using a consistent taxonomy makes it possible to correlate infrastructure, application, and user experience signals within a single platform.
Event lifecycle
Section titled “Event lifecycle”An event progresses through a lifecycle from creation to archival.
Events are ingested, validated, normalized, correlated with related events, and stored for analysis. Incidents may be created or updated as new events arrive.
This lifecycle ensures that raw monitoring data becomes actionable operational intelligence.
Correlation
Section titled “Correlation”Correlation connects events that are likely related to the same operational problem.
For example, a database latency spike, API health check failures, and checkout synthetic request failures may all be linked to a single database incident.
Correlation reduces noise and helps identify root causes more quickly.
Deduplication
Section titled “Deduplication”Deduplication prevents repeated observations from creating unnecessary alerts or incidents.
Multiple identical failures within a short period are grouped together, allowing operators to focus on meaningful changes rather than repetitive events.
This significantly reduces alert fatigue during ongoing outages.
Incident clustering
Section titled “Incident clustering”Incident clustering groups related events across services and dependencies into a single incident.
Instead of creating separate incidents for every failing health check, xrelia identifies the underlying disruption and creates a unified incident with affected services, dependencies, and timelines.
Clustering provides a clearer operational picture and simplifies incident response.