Prometheus SQL

Service that generates basic metrics for SQL result sets and exposing them as Prometheus metrics.

This service relies on the SQL Agent service to execute and return the SQL result sets.

Behavior

• Static configuration files are used to define the queries to monitor.
• Each query has a designated worker for execution.
• An interval is used to define how often to execute the query.
• Failed queries are automatically retried using a backoff mechanism.
• Faceted metrics are supported.
• A single metric's different facets can be filled in from different data sources.

Format

• Metric names are exposed in the format query_result_<metric name>.
• With faceted metrics, the name of the data column is determined by the data-field key in config, and all other columns (and column values) are exposed as labels.
• If the result set consists of a single row and column, the metric value is obvious and data-field is not needed.
• Label names under the same metric should be consistent.
• Each different query (query entry in config) for the same metric should lead to different label values.

Usage

Usage of prometheus-sql:
  -config string
        Configuration file to define common data sources etc.
  -host string
        Host of the service.
  -port int
        Port of the service. (default 8080)
  -queries string
        Path to file containing queries. (default "queries.yml")
  -queryDir string
        Path to directory containing queries.
  -service string
        Query of SQL agent service.

Queries file

A queries file is required for the application to know which data source to query and which queries that shall be monitored.

In the repository there is an example file that you can have a look at.

Config file

The config file is optional and can defined some default values for queries and data sources which can be referenced by queries. The benefit of referencing a data source will be reduction of duplication of database connection information. See example config file here and queries file which utilizes the config information.

Run via console

Create a queries.yml file in the current directory and run the following:

prometheus-sql

or for an alternate path, use the -queries or the -queryDir option:

prometheus-sql -queries ${PWD}/queries.yml

Run using Docker

Run the SQL agent service.

docker run -d --name sqlagent dbhi/sql-agent

Run this service. Mount the queries.yml file and link the SQL Agent service.

docker run -d \
  --name prometheus-sql \
  -p 8080:8080 \
  -v ${PWD}/queries.yml:/queries.yml \
  --link sqlagent:sqlagent \
  dbhi/prometheus-sql

If you want to separate database connection information etc you can do that by specifying data sources in separate file which you then can mount:

docker run -d \
  --name prometheus-sql \
  -p 8080:8080 \
  -v ${PWD}/queries.yml:/queries.yml \
  -v ${PWD}/prometheus-sql.yml:/prometheus-sql.yml \
  --link sqlagent:sqlagent \
  dbhi/prometheus-sql \
  -service http://sqlagent:5000 \
  -config prometheus-sql.yml

To view a plain text version of the metrics, open up the browser to the http://localhost:8080/metrics (or http://192.168.59.103:8080/metrics for boot2docker users).

Run using a Docker Compose file

Alternately, use the docker-compose.yml file included in this repository. The volumes section be added for mounting the queries.yml file.

Contributing

Read instructions how to contribute before you start.