Hostname - possibility to identify actual CF's app instance?

[krismarc]

Hey there,

Seems like CF's api is unaware of actual hostnames of application processes.

Wouldn't it be beneficial to have hostname included in eg stats? So we can easily identify instance to eg. restart specific instance only.

Since we know host IP and other container attributes, this wouldn't probably cost any additional processing time, no?

cf curl /v3/processes/94f9a32a-8c89-4d6d-bcac-fcd9b489c20c/stats | jq -r .
{
  "resources": [
    {
      "type": "web",
      "index": 0,
      "state": "RUNNING",
      "routable": true,
      "host": "100.106.178.59",
      "instance_internal_ip": "11.254.97.6",
      "uptime": 873203,
      "mem_quota": 2147483648,
      "disk_quota": 10737418240,
      "log_rate_limit": -1,
      "fds_quota": 16384,
      "isolation_segment": null,
      "details": null,
      "instance_ports": [
        {
          "external": 0,
          "internal": 8080,
          "external_tls_proxy_port": 61002,
          "internal_tls_proxy_port": 61001
        },
        {
          "external": 0,
          "internal": 8080,
          "external_tls_proxy_port": null,
          "internal_tls_proxy_port": 61443
        },
        {
          "external": 0,
          "internal": 2222,
          "external_tls_proxy_port": 61003,
          "internal_tls_proxy_port": 61002
        }
      ],
      "usage": {
        "time": "2025-06-04T10:12:53+00:00",
        "cpu": 0.006661531415403901,
        "cpu_entitlement": 0.07833967451462209,
        "mem": 639707986,
        "disk": 1208496128,
        "log_rate": 1073
      }
    },
    {
      "type": "web",
      "index": 1,
      "state": "RUNNING",
      "routable": true,
      "host": "100.106.210.23",
      "instance_internal_ip": "11.249.24.8",
      "uptime": 886817,
      "mem_quota": 2147483648,
      "disk_quota": 10737418240,
      "log_rate_limit": -1,
      "fds_quota": 16384,
      "isolation_segment": null,
      "details": null,
      "instance_ports": [
        {
          "external": 0,
          "internal": 8080,
          "external_tls_proxy_port": 61007,
          "internal_tls_proxy_port": 61001
        },
        {
          "external": 0,
          "internal": 8080,
          "external_tls_proxy_port": null,
          "internal_tls_proxy_port": 61443
        },
        {
          "external": 0,
          "internal": 2222,
          "external_tls_proxy_port": 61008,
          "internal_tls_proxy_port": 61002
        }
      ],
      "usage": {
        "time": "2025-06-04T10:12:53+00:00",
        "cpu": 0.004324711435490253,
        "cpu_entitlement": 0.0508586487237521,
        "mem": 555906599,
        "disk": 1151676416,
        "log_rate": 0
      }
    }
  ]
}

[krismarc]

I just had a brief look at the code:

The isntance guid seems to be known already?

cloud_controller_ng/lib/cloud_controller/diego/reporters/instances_stats_reporter.rb

Line 79 in b328c8f

instance_guid: actual_lrp.actual_lrp_instance_key.instance_guid,

and it's just not presented?

cloud_controller_ng/app/presenters/v3/process_stats_presenter.rb

Line 41 in b328c8f

instance_internal_ip: stats[:stats][:net_info][:instance_address],

[krismarc]

ah, seems like this was added in 3.191 😃 and we are not yet there..
f50766e

[philippthun]

This is resolved since API version 3.191 via PR #2831.

The GET /v3/processes/:guid/stats response now includes an instance_guid field containing the Diego instance identifier, which uniquely identifies each app instance and can be used to target specific instances (e.g. for restart).