06-dashboard.yaml

---
- hosts: pakupi_controller
  tasks:
  # TODO Pyslurm depends on Python 3.9 Ubuntu ships with 3.10
  # - name: Install required Python dependencies
  #  become: yes
  #  package:
  #    name:
  #    - python3-pip
  #    - cython3
  #    state: present
  # - name: Install Pyslurm library
  #  become: yes
  #  pip:
  #      name: git+http://github.com/PySlurm/pyslurm.git@v21.08.4
  
  # Install and configure Prometheus Slurm exporter
  - name: Install Go language support
    become: yes
    package:
      name:
      - golang
      state: present
  - name: Setup slurm exporter
    block:
    - name: Create temporary build directory
      tempfile:
        state: directory
        suffix: slurm
      register: pse_builddir
    - name: Download prometheus slurm exporter
      git:
        repo: https://github.com/vpenso/prometheus-slurm-exporter.git
        dest: "{{ pse_builddir.path }}/prometheus-slurm-exporter"
        version: '0.20'
    - name: "Fix unknown variables on build (Gaudeval/pakupi/issues/14)"
      lineinfile:
        path: "{{ pse_builddir.path }}/prometheus-slurm-exporter/Makefile"
        regexp: '^SHELL :='
        line: 'SHELL := $(shell which bash) -e -o pipefail'
    - name: Build slurm exporter
      make:
        chdir: "{{ pse_builddir.path }}/prometheus-slurm-exporter"
    - name: Install slurm exporter
      become: yes
      copy:
        dest: /usr/local/bin/prometheus-slurm-exporter
        src: "{{ pse_builddir.path }}/prometheus-slurm-exporter/bin/prometheus-slurm-exporter"
        remote_src: yes
        owner: root
        group: root
        mode: 'u=rwx,g=rwx,o=rx'
  - name: Create service location
    become: yes
    file:
      path: /usr/local/lib/systemd/system
      state: directory
      owner: root
      group: root
      mode: 'u=rwx,g=rwx,o=rx'
  - name: Install slurm exporter service
    become: yes
    copy:
      dest: /usr/local/lib/systemd/system/prometheus-slurm-exporter.service
      src: ./files/prometheus-slurm-exporter.service
      owner: root
      group: root
      mode: 'u=rw,g=r,o=r'
  - name: Start and enable slurm exporter service
    become: yes
    service:
      name: prometheus-slurm-exporter
      state: restarted
      enabled: yes
  # TODO Only restart service on definition change / service binary change
  # Install and configure Prometheus
  - name: Install prometheus
    become: yes
    package:
      name:
      - prometheus
      state: present
  - name: Add slurm exporter to Prometheus sources
    become: yes
    blockinfile:
      path: /etc/prometheus/prometheus.yml
      insertafter: '^scrape_configs:'
      block: |2
          - job_name: 'slurm'
            scrape_interval:  30s
            scrape_timeout:   30s
            static_configs:
            - targets: ['localhost:10100']
      validate: promtool check config %s
  - name : Restart prometheus
    become: yes
    service:
      name: prometheus
      state: restarted # Check config with http://<ip>:9090/classic/graph 
      enabled: yes
      
  # Install and configure grafana
  - name: Add Grafana repository key
    become: yes
    apt_key:
      url: https://packages.grafana.com/gpg.key
      state: present
  - name: Add Grafana repository
    become: yes
    apt_repository:
      repo: "deb https://packages.grafana.com/oss/deb stable main"
      state: present
      filename: grafana
  - name: Fetch packages from new repository
    become: yes
    apt:
      name: grafana
      state: present
      update_cache: yes
  - name: start Grafana server
    become: yes
    service:
      name: grafana-server
      state: started
      enabled: yes
  # TODO Set admin user and password https://grafana.com/docs/grafana/latest/setup-grafana/configure-grafana/#admin_user
  # TODO Configure default dashboard (json) https://grafana.com/docs/grafana/latest/setup-grafana/configure-grafana/#default_home_dashboard_path