Agent Skills: E2E Target VM

E2E Target VM

You are investigating an alert from an end-to-end test environment. The following infrastructure is reachable on the e2e-net Docker network:

• Target VM: hostname target (container e2e-target)
  • node_exporter metrics: http://target:9100/metrics
  • Demo HTTP service: http://target:8080/
• Prometheus: http://prometheus:9090
  • Instant query: http://prometheus:9090/api/v1/query?query=...
  • Range query: http://prometheus:9090/api/v1/query_range?query=...
• Alertmanager: http://alertmanager:9093
  • Active alerts: http://alertmanager:9093/api/v2/alerts

Investigation Cheatsheet

# Current CPU utilization (%)
curl -s 'http://prometheus:9090/api/v1/query?query=100-avg(rate(node_cpu_seconds_total{mode="idle"}[1m]))*100'

# Current memory utilization (%)
curl -s 'http://prometheus:9090/api/v1/query?query=(1-node_memory_MemAvailable_bytes/node_memory_MemTotal_bytes)*100'

# Current load average (1m)
curl -s 'http://prometheus:9090/api/v1/query?query=node_load1'

# Disk read throughput
curl -s 'http://prometheus:9090/api/v1/query?query=rate(node_disk_read_bytes_total[1m])'

# Confirm node_exporter is up
curl -s http://target:9100/metrics | head -20

When you have enough evidence, produce an RCA finding with a confidence score. The expected root cause for this test is synthetic CPU/memory stress generated by a stress-ng workload in the e2e-load-gen container.