PostgreSQL

Auto-Discovery

Health Check

Metrics

Logs

To enable metrics gathering, credentials are required. The agent will find the credentials by itself if PostgreSQL is running in a Docker container and username/password are set through the environment variables POSTGRES_USER (which defaults to postgres) and POSTGRES_PASSWORD.

By default, only sum metrics are gathered, if you want to monitor specific databases you can add them to the detailed_items setting.

If some auto-detected parameters are wrong, or you want to monitor specific databases, you can add the following to /etc/glouton/conf.d/99-local.conf:

service:
  [...]
  # For a PostgreSQL running outside a container
  - type: "postgresql"
    username: "USERNAME"
    password: "PASSWORD"
    address: "127.0.0.1"
    port: 5432

  # For an additional PostgreSQL running outside a container
  - type: "postgresql"
    instance: "NAME_ASSOCIATED_WITH_YOUR_ADDITIONAL_SERVICE"
    username: "USERNAME"
    password: "PASSWORD"
    address: "127.0.0.1"
    port: 5433

  # For a PostgreSQL running in a Docker container.
  - type: "postgresql"
    instance: "CONTAINER_NAME"
    username: "USERNAME"
    password: "PASSWORD"
    address: "172.17.0.2"
    port: 5432
    # Monitor the databases called "bleemeo" and "postgres".
    detailed_items:
      - bleemeo
      - postgres

When using Docker, you may use labels to set the databases to monitor and the credentials:

docker run --label glouton.detailed_items="bleemeo,postgres" --label glouton.username="USERNAME" --label glouton.password="PASSWORD" [...]

Agent gathers the following metrics:

Metric	Description
`service_status`	Status of PostgreSQL
`postgresql_blk_read_utilization`	PostgreSQL reading data file blocks utilization
`postgresql_blk_read_utilization_sum`	Sum of PostgreSQL reading data file blocks utilization
`postgresql_blk_write_utilization`	PostgreSQL writing data file blocks utilization
`postgresql_blk_write_utilization_sum`	Sum of PostgreSQL writing data file blocks utilization
`postgresql_blks_hit_sum`	Number of blocks read from PostgreSQL cache per second
`postgresql_blks_read_sum`	Number of blocks read from disk per second
`postgresql_commit_sum`	Number of commits per second
`postgresql_rollback_sum`	Number of rollbacks per second
`postgresql_temp_bytes_sum`	Temporary file write throughput in bytes per second
`postgresql_temp_files_sum`	Number of temporary files created per second
`postgresql_tup_deleted_sum`	Number of rows deleted per second
`postgresql_tup_fetched_sum`	Number of rows fetched per second
`postgresql_tup_inserted_sum`	Number of rows inserted per second
`postgresql_tup_returned_sum`	Number of rows returned per second
`postgresql_tup_updated_sum`	Number of rows updated per second

Except for the status metric, all metrics also exist without the _sum suffix if detailed_items is used. For instance, the metric postgresql_commit with the item mycontainer_mydb corresponds to the number of commits per second on the database mydb running in the container mycontainer.