What Happens During an Incident (Part 4)

Short summary

Incident investigation requires separating signals—metrics reveal where failures occur, logs/traces expose what happened, audit proves causality. For global systems, correlate by region and tenant to distinguish customer-specific from systemic failures; AWS DevOps Agent automates telemetry correlation and investigation timelines. Because production tracing is sampled, investigate similar requests in the same time window when exact traces don't exist.

• •Separate signals: metrics show location, logs/traces show what happened, audit proves causality
• •For global systems, use correlation IDs and regional boundaries to distinguish customer-specific from systemic failures
• •AWS DevOps Agent automates telemetry correlation and investigation—critical when production tracing is sampled

Generated with AI, which can make mistakes.