Ansible Role: Prometheus

Prometheus is an opensource monitoring solution that gathers time series based numerical data. It is a project which was started by Google’s ex-employees at SoundCloud.

To monitor your services and infra with prometheus your service need to expose an endpoint in the form of port or url. For ex:- {{ localhost:9090 }}. The endpoint is HTTP interface that exposes the metrics.

This role is compatible with Linux Hardened OS

Version History

Date	Version	Description	Changed By
Feb 2019	v0.0.1	Initial Draft	Abhishek Dubey
July 2020	v0.1.0	Added Service Discovery Feature	Mahesh Kumar
April 2022	v0.1.0	Added ec2 sd configs plugin	Varghese Kurian

Requirements

There is no particular requirment for running this role. As this role is platform independent for centos 6 or 7 and ubuntu 14, 16 & 18. The only dependency for centos 6 is libselinux python and we have included that as well. The basic requirments are:-

• Centos/Ubuntu Server
• Python should be installed on the target server so that ansible can perform task
• libselinux-python should be available on Centos 6 so that ansible can connect to it

Role Variables

# defaults file for prometheus
prometheus_version: "2.3.2"
prometheus_ip: "0.0.0.0"
prometheus_port: "9090"
base_download_url: "https://github.com/prometheus/prometheus/releases/download"
user_name: "prometheus"
group_name: "prometheus"
prometheus_binary_path: "/usr/local/bin"

# please enter telegraf or node_exporter
agent_discovery: "no"
agent: "telegraf"
tgraf_port: "9273"
node_exporter_port: "9100"
env: prod

# please enter ec2 config details to enable
enable_ec2_sd_configs: "no"
aws_region: us-east-1
filter_tag_key: NodeExporter #Tag-Key to filter EC2 nodes
filter_tag_value: Enable     #Tag-Value to filter EC2 nodes

You can define any prometheus version that you want to install on your server.

Mandatory Variables

Variable	Default Value	Possible Values	Description
prometheus_version	"2.3.2"	Any Version	Prometheus will be downloaded from github releases, so you have to define version
prometheus_ip	"0.0.0.0"		IP of the server on which prometheus will expose its UI

Optional Variables

Variable	Default Value	Possible Values	Description
prometheus_port	"9090"	Prometheus Service Port	Port no. of server on which prometheus should listen
base_download_url	"https://github.com/prometheus/prometheus/releases/download"	Base url of Prometheus's Download link	Base url of prometheus release
user_name	prometheus	Any User	Prometheus User Name
group_name	prometheus	Any Group	Promtheus Group Name
prometheus_binary_path	"/usr/loca/bin"	Any Path	Path to Copy Prometheus Binary
agent_discovery	"no"	"yes"/"no"	Enable of Disable Agent Discovery, Agent Discovery added all the agents in Promtheus Configuration
agent	"telegraf"	"telegraf"/"node_exporter"	Which agent you are using to get metrics. Supported Values are "node_exporter" and "telegraf"
env	"prod"	Any Enviornment Name	Enviornment, in which you agents are running
tgraf_port	"9273"	telegraf service port	if you are using telegraf, Port no. of telegraf on which telegraf is listening
node_exporter_port	"9100"	Node Exporter Service Port	if Node Exporter is using as agent, Port no. of telegraf on which telegraf is listening
Enable_ec2_sd_configs	"yes"	yes or no	If you want to enable the ec2 plugin service then it should be yes
Filter_tag_key	NodeExporter	Can be any name	Tag should be exporter name for better understanding
Filter_tag_value	Enable	Can be any Bulien value	value should be bulien other than that ec2 plugin will no support

Dependencies

• If Service Discovery is Enable, install one of telgraf and node exporter agent on each node, which must be added in prometheus targets.

Directory Structure

Directory structure:-

osm_prometheus
├── defaults
│   └── main.yml
├── handlers
│   └── main.yml
├── hosts
├── README.md
├── site.yml
├── tasks
│   ├── debian.yml
│   ├── main.yml
│   └── redhat.yml
└── templates
    ├── prometheus.init.j2
    ├── prometheus.service.j2
    └── prometheus.yml.j2

Example Playbook if Agent Discovery is not Enable

Here is an example for the main playbook

---
- hosts: prometheus
  user: test-user
  become: true
  roles:
    -  prometheus

For inventory you can create a host file in which you can define your server ip, For example:-

[prometheus]
10.1.1.100

Example Playbook if Agent Discovery is Enable

Here is an example for the main playbook. Following Playbook Gather Facts of each Node, where agents is installed. Because of Gather Facts, Ansible Controller storing ip address of each host and adding it in promtheus configuration.

---
- name: Gather Facts of Agent Nodes
  hosts: agents
  gather_facts: true
  tasks: []

- hosts: prometheus
  user: test-user
  become: true
  roles:
    -  prometheus

For inventory you can create a host file in which you can define your server ip, For example:-

[prometheus]
10.1.1.100

[agents]
10.1.1.100
10.1.1.101
10.1.1.102

You can simply use this role by using this command

ansible-playbook -i hosts site.yml

New Feature requests

Dynamic discovery support

• AWS
• Azure
• DNS

License

GNU

Author Information

This role is written and maintained by Abhishek Dubey. If you have any queries and sugesstion, please feel free to reach.