01 / TELEMETRY

Collect what joins.

Edge and service

Request timestamp, host, route, status, bytes, source and upstream instance. Respect application privacy when recording headers and bodies.

JVM and OS

Application exceptions, JVM fatal-error reports, process exits or restarts, CPU and memory changes, and service logs.

Process and network

Unexpected children of Java processes, outbound destinations, file changes and identity or privilege changes.

02 / TIMELINE

Follow one instance over time.

Join request and JVM events by service, host, container and time. A spike in failures followed by a worker restart or unusual child process deserves investigation; each signal also has benign causes.

Request anomaly→JVM exception or restart→Host activity
03 / TRIAGE

Retain evidence before changing the service.

Preserve relevant request logs, service logs, crash reports, process trees and deployment metadata under your incident process. Compare against normal jobs and maintenance activity. A quiet dashboard does not establish safety.

Read the current exploit evidence status →