Question posted in 
I am trying to use Prometheus Role: https://github.com/prometheus-community/ansible
I created a playbook – playbook.yaml, with the following content:
---
- hosts: prometheus-server
become: true
roles:
- role: prometheus.prometheus.prometheus
prometheus_version: "2.43.0"
prometheus_config_dir: "/sites/prometheus"
prometheus_web_listen_address: "0.0.0.0:9090"
prometheus_global: {"evaluation_interval": "1m5s", "scrape_interval": "1m", "scrape_timeout": "59s"}
prometheus_external_labels: {"environment": "env1"}
prometheus_alert_rules_files: ["/sites/prometheus/rules.d/rules.yml"]
prometheus_alertmanager_config:
- scheme: http
static_configs:
- targets:
- alertmanager-1.xxx:9093
The inventory is OK:
all:
hosts:
prometheus-server:
ansible_host: x.x.x.x
vars:
ansible_user: ubuntu
ansible_ssh_private_key_file: ~/.ssh/id_rsa_test
ansible_ssh_common_args: '-o ProxyCommand="ssh -W %h:%p -i ~/.ssh/id_rsa [email protected]"'
I changed all these values to use rules.d:
In roles/prometheus/templates/prometheus.yml.j2:
{% if not prometheus_agent_mode and prometheus_alert_rules_files != [] %}
rule_files:
- {{ prometheus_config_dir }}/rules.d/*.rules
{% endif %}
In roles/prometheus/tasks/configure.yml
src: "{{ item }}"
dest: "{{ prometheus_config_dir }}/rules.d/"
owner: root
In ansible/roles/prometheus/meta/argument_specs.yml, nothing to change:
prometheus_alert_rules:
description:
- "Full list of alerting rules which will be copied to C({{ prometheus_config_dir }}/rules.d/ansible_managed.rules)."
- "Alerting rules can be also provided by other files located in C({{ prometheus_config_dir }}/rules.d/) which have C(*.rules) extension"
- "Please see default values in role defaults/main.yml"
type: "list"
elements: "dict"
When running the playbook, it still returns the default route (even if changed and saved).
I also tried manually adding other tasks, that are not being executed, in tasks:
But nothing…
In my server, I manually changed the value and restarted prometheus, to run the playbook and check that it changes again to the "rules" route (to see that the playbook works).
Not sure what I can do…
To help with troubleshooting:
- Permissions in the destination server are the same.
- I even created some tasks that are not being executed in the role.
- SCP WARNINGS are triggered in the playbook (I think that is not related to the problem).
Some of the warnings I receive:
Adding WARNINGS:
[WARNING]: sftp transfer mechanism failed on [X.X.X.X]. Use ANSIBLE_DEBUG=1 to see detailed information
[WARNING]: scp transfer mechanism failed on [X.X.X.X]. Use ANSIBLE_DEBUG=1 to see detailed information
2
Answers
The problem was finally related to using prometheus.prometheus.prometheus, and because the playbook was inside of the role, causing some problems while executing.
By changing the route of the playbook, and using only "prometheus" as role, this was solved.
If this is the exact content of your playbook, your variables are ignored because you passed them outside of the
varskeyword (please check Using roles at the play level for details). It should have been something like this:Alternatively, and I would call it a preferable way as it’s more flexible and doesn’t require playbook changes, you could put the variables directly into your inventory: