Move from “origin is red on the BCS scoreboard” to the failing dependency and span—without treating APM as the only end-to-end model.
Use after BCS health and async ES|QL bridges implicate application hops. This pack ships a runbook plus ES|QL for dependency errors, slow spans on failures, entity-scoped traces, and service failure rates. Hop-local trace.id is the drill-down; session/content/message keys remain the cross-layer spine.
Continue into Kubernetes Observability when capacity is implicated, or Network Observability when the path is guilty and app spans are clean.
What’s included
- Investigation runbook: BCS → entity filter → dependencies → waterfall → infra/network branch
- ES|QL query set for APM/OTel traces
- OTel attribute reminders (
http.route,db.system,messaging.system)
Source: Elastic APM
| Tested versions | |
| ECS compliant |
You must log in to submit a review.