Introducing Correlations: from symptom to system state

Correlations in Axiom streamline the process of identifying and resolving issues by connecting related telemetry data, allowing engineers to seamlessly navigate from initial symptoms to the comprehensive system state without manual query efforts. When an API times out, for instance, engineers can start with a log event, which serves as the first clue, and then use Correlations to access the corresponding trace without switching data sources or manually copying trace IDs. This trace reveals the request's journey through the system, highlighting involved services and potential latency sources, thereby narrowing down the investigation. Moreover, Correlations enable the inspection of workload metrics to determine if a single request issue is indicative of a broader service problem, as simultaneous access to these metrics can reveal if the service was under pressure from rising latency, throughput changes, or resource saturation, helping to distinguish between isolated incidents and larger patterns.

No tracked trend matches for this post yet.