We require the ability to customize the deployed Node Exporter collectors within the EDPM Telemetry configuration for critical business monitoring needs.

Currently, the collector list is hardcoded in the EDPM Ansible role ([1]), which prevents us from gathering specific metrics essential for our environment. 

 [1]https://github.com/openstack-k8s-operators/edpm-ansible/blob/2f8404c86d0879021df8bb5b4ef5e28dd9c164fc/roles/edpm_telemetry/templates/node_exporter.json.j2#L8

Bug impact

• Impact: No possibility to gather required metrics for business operations and performance monitoring.

Known workaround

Customer have implemented a temporary workaround by creating a custom OpenStackDataPlaneService. This service overrides the default Node Exporter deployment with a new container that includes our required collectors (e.g., specific systemd units, sysctl includes, and collector disabling).

• apiVersion: dataplane.openstack.org/v1beta1
  kind: OpenStackDataPlaneService
  metadata:
    name: xxxx-node-exporter-service
    namespace: openstack
  spec:
    playbookContents: |
      - hosts: all
        become: true
        tasks:
          - name: Prepare textfile directory
            ansible.builtin.file:
              state: directory
              path: /var/lib/node_exporter/textfile_collector/
              recurse: true
              mode: '0755'
              owner: root
              group: root
        
          # Based on upstream https://github.com/openstack-k8s-operators/edpm-ansible/blob/main/roles/edpm_telemetry/templates/node_exporter.json.j2 
          # Modified for XXXXX needs
          - name: Write node_exporter.json config file
            ansible.builtin.copy:
              dest: /var/lib/openstack/config/telemetry/node_exporter.json
              mode: '0644'
              content: |
                {
                    "image": "{{ edpm_telemetry_node_exporter_image }}",
                    "restart": "always",
                    "recreate": true,
                    "user": "root",
                    "privileged": true,
                    "ports": ["9100:9100"],
                    "command": [
                        "--web.config.file=/etc/node_exporter/node_exporter.yaml",
                        "--collector.systemd",
                        "--collector.systemd.unit-include=(edpm_.*|ovs.*|openvswitch|virt.*|rsyslog|auditbeat|qualys-cloud-agent|puppet)\\.service",
                        "--web.disable-exporter-metrics",
                        "--collector.netclass.ignored-devices=^(lo|tap.+|genev.+|ovs.+|br.+)$",
                        "--collector.netdev.device-exclude=^(lo|tap.+|genev.+|ovs.+|br.+)$",
                        "--collector.sysctl",
                        "--collector.sysctl.include=fs.aio-nr",
                        "--collector.sysctl.include=fs.aio-max-nr",
                        "--collector.textfile.directory=/var/lib/node_exporter/textfile_collector",
                        "--no-collector.cpufreq",
                        "--no-collector.entropy",
                        "--no-collector.fibrechannel",
                        "--no-collector.infiniband",
                        "--no-collector.thermal_zone",
                        "--no-collector.time",
                        "--no-collector.timex",
                        "--no-collector.stat",
                        "--no-collector.hwmon",
                        "--no-collector.selinux",
                        "--no-collector.powersupplyclass",
                        "--no-collector.pressure",
                        "--no-collector.rapl"
                    ],
                    "net": "host",
                    "environment": {
                        "OS_ENDPOINT_TYPE":"internal"
                    },
                    "healthcheck": {
                        "test": "/openstack/healthcheck node_exporter",
                        "mount": "/var/lib/openstack/healthchecks/node_exporter"
                    },
                    "volumes": [
                        "/var/lib/openstack/config/telemetry/node_exporter.yaml:/etc/node_exporter/node_exporter.yaml:z",
                        "/var/lib/openstack/certs/telemetry/default:/etc/node_exporter/tls:z",
                        "/var/run/dbus/system_bus_socket:/var/run/dbus/system_bus_socket:rw",
                        "/var/lib/node_exporter/textfile_collector:/var/lib/node_exporter/textfile_collector:ro,z"
                    ]
                }
  
          - name: Deploy node-exporter container using upstream method
            ansible.builtin.include_role:
              name: osp.edpm.edpm_container_manage
            vars:
              edpm_container_manage_config: "/var/lib/openstack/config/telemetry"
              edpm_container_manage_healthcheck_disabled: true
              edpm_container_manage_clean_orphans: false
              edpm_container_manage_config_patterns: "node_exporter.json"
            
          - name: Restart node-exporter container
            become: true
            ansible.builtin.systemd:
              state: restarted
              name: "edpm_node_exporter.service"