sql_exporter

sql_exporter

Overview

sql_exporter generates basic metrics for SQL result sets and expose them as Prometheus metrics.

Pros:

• Configuration driven application
• Logger included for debugging purpose
• Support integer and float metrics
• Each query can be run at a different time interval

Cons:

• Currently, only support the gauge metric type
• Only support SQL databases

Usage

The depot contain a go.mod and a go.sum files. Build the binary by running the following command:

$ go build

Then run it from the command line:

$ ./sql_exporter

By default, the path to the configuration file is ./config.yaml. You can change it using the -config flag:

$ ./sql_exporter -config myconfiguration.yaml

Use the -help flag to get help information:

Usage of ./sql_exporter:
  -config string
        path to the configuration file (default "config.yaml")
  -help
        display the help message

Configuration

Here is a sample configuration describing the default values with comments:

# The address on which the server will list (interface:port)
addr: ":8080"

# One entry per database you want to connect to. Databases 
# and metrics are handled in parallel.
databases:

#    # Address of the database (address:port)
#  - address: localhost:3306
#    user: web_admin
#    password:
#    name: website_information

#    # List of metrics to retrieve for this database
#    metrics:

#        # Statement to execute
#      - statement: "select count(visit_id) from visits"
#        # Interval at which the statement is executed. If the
#        # statement takes more than the interval duration to
#        # run, the next execution will start right away.
#        interval: 60m
#        # Name of the metric to expose.
#        name: "namespace_subsystem_name_unit"
#        # Help message of the metric.
#        help: "number of visits on the website"
#        # Labels to add to the metric.
#        labels:
#          data_source: "analytics"

sql_exporter as a service

To expose your metrics continuously, you can run sql_exporter as a service. You can use sql_exporter.service as a template to create your own service configuration:

# /usr/lib/systemd/system/sql_exporter.service
[Unit]
Description=SQL Exporter, Prometheus exporter that collect metrics from SQL databases
Documentation=https://github.com/rmasciulli/sql_exporter/blob/master/README.md
After=network-online.target

[Service]
Type=simple
ExecStart=/opt/sql_exporter/sql_exporter -conf config.yaml
KillSignal=SIGINT
KillMode=process
Restart=on-failure

[Install]
WantedBy=multi-user.target