Add node name in the container metrics because of system_process

[marceloamaral]

The issue #365 identified that the system_process "container" energy consumption is being aggregated across all nodes.
This is because the name and namespace are the same in all nodes.

To fix that, we can introduce the node label in the container metrics

Signed-off-by: Marcelo Amaral marcelo.amaral1@ibm.com

[rootfs]

each metric emits the instance name, e.g.

we don't need to add addtional node name in the metric

{
  "__name__": "kepler_pod_energy_stat",
  "block_devices_used": "1",
  "command": "python3.8",
  "container": "kepler-exporter",
  "container_name": "server-api",
  "curr_bytes_read": "0",
  "curr_bytes_writes": "0",
  "curr_cache_miss": "87188",
  "curr_container_cpu_usage_seconds_total": "0",
  "curr_container_memory_working_set_bytes": "0",
  "curr_cpu_cycles": "51576400",
  "curr_cpu_instr": "67433499",
  "curr_cpu_time": "6",
  "endpoint": "http",
  "instance": "xxxxx",  <--- this is the node name
  "job": "kepler-exporter",
  "namespace": "kepler",
  "pod": "kepler-exporter-hp5x2",
  "pod_name": "kepler-model-server-68b9ccbfd6-xsjps",
  "pod_namespace": "monitoring",
  "service": "kepler-exporter",
  "total_bytes_read": "0",
  "total_bytes_writes": "1236992",
  "total_cache_miss": "6316520376",
  "total_container_cpu_usage_seconds_total": "3462",
  "total_container_memory_working_set_bytes": "47489024",
  "total_cpu_cycles": "4008660406685",
  "total_cpu_instr": "5408344083710",
  "total_cpu_time": "52555158"
}

[marceloamaral]

You are right...sorry for the mistake.

kepler_container_joules_total{container="kepler-exporter", container_name="kube-rbac-proxy", 
container_namespace="monitoring", endpoint="http", instance="kind-control-plane", job="kepler-exporter", 
namespace="kepler", pod="kepler-exporter-65b6m", pod_name="prometheus-operator-7b64d465b9-2598k", 
service="kepler-exporter"}