Hi - I use your excellent role to install the node_exporter to my Linux machines. Would you welcome PR(s) that add Windows (2016 Server) support?

That would look like:

• generate a new user for the WMI exporter to run as
• make that user a local administrator (because the permissions required are byzantine, see discussion inside prometheus-community/windows_exporter#233)
• download a given version of the WMI exporter, unzip, configure
• install NSSM to turn the WMI exporter into a service with a minimum of hassle
• make the service log to a file
• assert success by fetching from the local /metrics endpoint

It would be nice to be able to configure Prometheus to support remote and write: https://prometheus.io/docs/prometheus/latest/configuration/configuration/#remote_write
https://prometheus.io/docs/prometheus/latest/configuration/configuration/#remote_read

This should fix it: #78

option to install the blackbox-exporter

I installed the collations with the command,

ansible-galaxy install ernestas-poskus.ansible-prometheus

ansible-galaxy list
/home/ubuntu/.ansible/roles
ernestas-poskus.ansible-prometheus, v4.18.0

I run the command as follows:
ansible-playbook -i hosts -l group1 playbook.yml --check

I got errors as follows:

ERROR! the role 'ansible-prometheus' was not found in /home/ubuntu/works/roles:/home/ubuntu/.ansible/roles:/usr/share/ansible/roles:/etc/ansible/roles:/home/ubuntu/works

The error appears to be in '/home/ubuntu/works/playbook.yml': line 5, column 7, but may
be elsewhere in the file depending on the exact syntax problem.

The offending line appears to be:

  roles:
    - role: ansible-prometheus
      ^ here

As follows is the playbook:

- name: Installing Prometheus on hosted machine
  hosts: group1
    #sudo: yes
  roles:
    - role: ansible-prometheus
      prometheus_config_scrape_configs:
        - job_name: 'prometheus'
          honor_labels: true
          scrape_interval: '15s'
          scrape_timeout: '3s'
          metrics_path: '/metrics'
          scheme: 'http'
          static_configs:
            - targets:
                - 'localhost:9090' # Prometheus itself
                - 'localhost:9100' # Node exporter
        - job_name: 'consul-services'
          consul_sd_configs:
            - server: "localhost:8500"

I tried to use this playbook to install prometheus version 2, but it's failing because the right syntax in version 2 is promtool check rules instead of promtool check-rules.
It happens the same for promtool check config.

E.g. "DOCS: Prometheus variables" in the README at https://galaxy.ansible.com/ernestas-poskus/ansible-prometheus points to a non-existent URL https://galaxy.ansible.com/docs/prometheus.md

icmp probe is returning failed

probe requires elevated privileges as outlined in the blackbox_exporter docs

when a playbook is run in check mode its failing when trying to elevate permissions of the install directory.

It fails as it has not downloaded the relevant install file so directory does not yet exist

therefore the task that sets the elevated directory permissions needs when: not check_mode needs to be added to task name: Apply elevated privileges to blackbox_exporter for ICMP probe in https://github.com/ernestas-poskus/ansible-prometheus/blob/master/tasks/install-blackbox_exporter.yml

//Added later
Hello!

During provisioning my Vagrant box with prometheus i have issue as below

Ansible 2.3.2.0 fails during installation of systemd service:
TASK [ernestas-poskus.ansible-prometheus : Install Prometheus systemd service] *** fatal: [default]: FAILED! => {"changed": false, "failed": true, "msg": "AnsibleUndefinedVariable: 'dict object' has no attribute 'iteritems'"}

The issue probably occurs because python 3 don't have attribute iteritems, on Ansible with python 2 role works as intended

https://wiki.python.org/moin/Python3.0 (abotu changes between versions)

Kind regards

in my playbook I have the following var set to define my receivers

prometheus_alert_manager_config_receivers:
  - name: 'default'
    slack_configs:
      - channel: '#operations'
        send_resolved: true
    email_configs:
      - to: 'user@emailaddress'
        send_resolved: true

when the playbook is run the config gets generated like this:

receivers:
-   email_configs:
    -   send_resolved: true
        to: user@emailaddress
    name: default
    slack_configs:
    -   channel: '#operations'
        send_resolved: true

I had a problem where the alert notification wasn't working, alerts we're showing up in prometheus, but not making it to alertmanager and hence the notification receivers were not working.

I've added the following manually to prometheus.yml and alert notifications is now working via alertmanager and receivers.

alerting:
  alertmanagers:
  - scheme: http
    static_configs:
    - targets:
      - "localhost:9093"

I've had a look at the role variables and can't seem to see any related variables to set these.

rules are not removed from the /etc/prometheus/rules directory if they are removed from role variables or are renamed.

Ok I'm not clear on how to define the slack receiver text and description as per https://prometheus.io/docs/alerting/notification_examples/#ranging-over-all-received-alerts for this role:

My config:

  - name: 'default'
    slack_configs:
      - channel: '#alerts'
        username: 'Alert-Manager'
        title: "[[ range $Alerts ]][[ $Annotations.summary ]]\n[[ end ]]"
        text: "[[ range $Alerts ]][[ $Annotations.description ]]\n[[ end ]]"
        send_resolved: true

What is being generated in prometheus.yml

    slack_configs:
    -   channel: '#alerts'
        send_resolved: true
        text: '[[ range $Alerts ]][[ $Annotations.description ]]

            [[ end ]]'
        title: '[[ range $Alerts ]][[ $Annotations.summary ]]

            [[ end ]]'

first it appears that the [[ ]] are not being converted to { } for the receiver fields

secondly I'm not sure if the I'm required to put the $ in front or leave as per the above link i.e. .Alerts and .Annotations.decription etc

just like the push-gateway install, an option to install the snmp-exporter with basic mib walk.

Prometheus: 2.4.3
Alert Manager: 0.15.2

Latest release of node-exporter is introducing breaking changes in metrics name. I think that it would be nice to add some note to README when you will bump node_exporter version. That way if someone does not follow node-exporter repository at least will have chance to not be surprised when everything stops working(alerts, dashboards).

Link to rc: https://github.com/prometheus/node_exporter/releases/tag/v0.16.0-rc.0