HomeAudit/playbooks/netdata-parent-child.yml

---
- name: Configure Netdata Parent-Child Streaming Architecture
  hosts: all
  become: yes
  vars:
    netdata_parent_host: "192.168.50.225"  # Fedora workstation wired IP (more stable)
    netdata_stream_api_key: "{{ netdata_api_key | default('homelab-stream-key-2024') }}"
    
  tasks:
    - name: Install Netdata on hosts where it's missing
      block:
        - name: Install Netdata via package manager (Fedora/RHEL)
          dnf:
            name: netdata
            state: present
          when: ansible_os_family == "RedHat"
          
        - name: Install Netdata via package manager (Debian/Ubuntu)
          apt:
            name: netdata
            state: present
            update_cache: yes
          when: ansible_os_family == "Debian"
          
        - name: Install Netdata via installer script (fallback)
          shell: |
            curl -L https://get.netdata.cloud/kickstart.sh | bash -s -- --stable-channel --disable-telemetry --dont-wait
          args:
            creates: /usr/sbin/netdata
          when: ansible_os_family not in ["RedHat", "Debian"]
      when: inventory_hostname in ['surface', 'fedora', 'omvbackup']

    - name: Configure Netdata Parent (Central monitoring node)
      block:
        - name: Create Netdata parent configuration
          copy:
            content: |
              [global]
                  # Parent node configuration
                  memory mode = dbengine
                  page cache size = 32
                  dbengine disk space = 2048
                  
                  # Network settings
                  bind to = *:19999
                  allow connections from = 192.168.50.0/24 100.64.0.0/10
                  
                  # Disable local data collection plugins to reduce load
                  [plugins]
                      apps = no
                      cgroups = no
                      diskspace = no
                      
              [web]
                  # Allow dashboard from any host on home network
                  allow dashboard from = 192.168.50.0/24 100.64.0.0/10
                  
              # Stream configuration for receiving child data
              [stream]
                  enabled = yes
                  destination =
                  api key = {{ netdata_stream_api_key }}
                  timeout seconds = 60
                  buffer size bytes = 1048576
                  reconnect delay seconds = 5
                  initial clock resync iterations = 60
                  
                  # Accept streams from child nodes
                  [{{ netdata_stream_api_key }}]
                      enabled = yes
                      allow from = 192.168.50.0/24 100.64.0.0/10
                      default history = 3600
                      default memory = dbengine
                      health enabled by default = auto
            dest: /etc/netdata/netdata.conf
            backup: yes
            
        - name: Enable and start Netdata parent
          systemd:
            name: netdata
            enabled: yes
            state: restarted
            
        - name: Open firewall for Netdata (firewalld)
          firewalld:
            port: 19999/tcp
            permanent: yes
            state: enabled
            immediate: yes
          ignore_errors: yes
          
        - name: Open firewall for Netdata (ufw)
          ufw:
            rule: allow
            port: '19999'
            proto: tcp
          ignore_errors: yes
          
      when: inventory_hostname == 'fedora'

    - name: Configure Netdata Children (streaming to parent)
      block:
        - name: Configure Netdata child to stream to parent
          copy:
            content: |
              [global]
                  # Child node - minimal local storage
                  memory mode = none
                  
                  # Network settings
                  bind to = localhost:19999
                  
              [web]
                  # Restrict dashboard access to localhost only
                  allow dashboard from = localhost
                  
              # Stream all data to parent
              [stream]
                  enabled = yes
                  destination = {{ netdata_parent_host }}:19999
                  api key = {{ netdata_stream_api_key }}
                  timeout seconds = 60
                  buffer size bytes = 1048576
                  reconnect delay seconds = 5
                  initial clock resync iterations = 60
                  
                  # Send everything to parent
                  send charts matching = *
                  
              # Reduce resource usage
              [plugins]
                  # Keep essential monitoring but reduce frequency
                  proc = yes
                  diskspace = yes
                  cgroups = yes
                  apps = yes
                  
              # Optimize collection intervals
              [plugin:proc]
                  /proc/net/dev = yes
                  /proc/diskstats = yes
                  /proc/net/sockstat = yes
                  /proc/meminfo = yes
                  /proc/vmstat = yes
                  /proc/stat = yes
                  /proc/loadavg = yes
                  
            dest: /etc/netdata/netdata.conf
            backup: yes
            
        - name: Restart Netdata child
          systemd:
            name: netdata
            state: restarted
            
      when: inventory_hostname != 'fedora'

    - name: Verify Netdata service status
      systemd:
        name: netdata
        state: started
        enabled: yes
      register: netdata_status
      
    - name: Display Netdata status
      debug:
        msg: |
          Netdata {{ 'parent' if inventory_hostname == 'fedora' else 'child' }} configured on {{ inventory_hostname }}
          Status: {{ netdata_status.status.ActiveState }}
          {% if inventory_hostname == 'fedora' %}
          Parent dashboard: http://{{ ansible_host }}:19999
          {% else %}
          Streaming to: {{ netdata_parent_host }}:19999
          {% endif %}

    - name: Test connectivity to parent (from children)
      uri:
        url: "http://{{ netdata_parent_host }}:19999/api/v1/info"
        method: GET
        timeout: 10
      register: parent_test
      ignore_errors: yes
      when: inventory_hostname != 'fedora'
      
    - name: Display connectivity test results
      debug:
        msg: |
          Connection to parent: {{ 'SUCCESS' if parent_test.status == 200 else 'FAILED' }}
          {% if parent_test.status != 200 %}
          Error: {{ parent_test.msg | default('Unknown error') }}
          {% endif %}
      when: inventory_hostname != 'fedora' and parent_test is defined

  handlers:
    - name: restart netdata
      systemd:
        name: netdata
        state: restarted