The easiest and recommended way to setup Grafana on Aptible is using the Aptible Metrics Terraform Module. This provisions Aptible metric drains with pre-built Grafana dashboards and alerts for monitoring RAM and CPU usage for your Aptible apps and databases. This simplifies the setup of metric drains so you can start monitoring your Aptible resources immediately, all hosted within your Aptible account. If you would rather set it up from scratch, use this guide.
Grafana is an open-source platform for analytics and monitoring. It's an ideal choice to use in combination with an InfluxDB metric drain.
Here's what makes Grafana so compelling:
- It makes it easy to build beautiful graphs and set up alerts.
- It works out of the box with InfluxDB.
- It works very well in a containerized environment like Aptible.
This guide will cover setting up Grafana on Aptible.
Provision a PostgreSQL database
Grafana needs a Database to store sessions and Dashboard definitions. It works great with PostgreSQL, which you can deploy on Aptible.
Configure the database
Once you have created the PostgreSQL Database, create a tunnel using the
aptible db:tunnel command, then connect using
psql and run the following commands to create a
sessions database for use by Grafana:
CREATE DATABASE sessions;
Then, connect to the newly-created
And finally, create a table for Grafana to store sessions in:
CREATE TABLE session ( key CHAR(16) NOT NULL, data BYTEA, expiry INTEGER NOT NULL, PRIMARY KEY (key) );
Deploy the Grafana app
Grafana is available as a Docker image and can be configured using environment variables. As a result, you can use Direct Docker Image Deploy to easily deploy Grafana on Aptible.
Here is the minimal deployment configuration to get you started. In the example below, you'll have to substitute a number of variables:
$ADMIN_PASSWORD: Generate a strong password for your Grafana
$SECRET_KEY: Generate a random string (40 characters will do).
$YOUR_DOMAIN: The domain name you intend to use to connect to Grafana (e.g.
$DB_USERNAME: The username for your PostgreSQL database. For a PostgreSQL database on Aptible, this will be
$DB_PASSWORD: The password for your PostgreSQL database.
$DB_HOST: The host for your PostgreSQL database.
$DB_PORT: The port for your PostgreSQL database.
aptible apps:create grafana aptible deploy --app grafana --docker-image grafana/grafana \ "GF_SECURITY_ADMIN_PASSWORD=$ADMIN_PASSWORD" \ "GF_SECURITY_SECRET_KEY=$SECRET_KEY" \ "GF_DEFAULT_INSTANCE_NAME=aptible" \ "GF_SERVER_ROOT_URL=https://$YOUR_DOMAIN" \ "GF_SESSION_PROVIDER=postgres" \ "GF_SESSION_PROVIDER_CONFIG=user=$DB_USERNAME password=$DB_PASSWORD host=$DB_HOST port=$DB_PORT dbname=sessions sslmode=require" \ "GF_LOG_MODE=console" \ "GF_DATABASE_TYPE=postgres" \ "GF_DATABASE_HOST=$DB_HOST:$DB_PORT" \ "GF_DATABASE_NAME=db" \ "GF_DATABASE_USER=$DB_USERNAME" \ "GF_DATABASE_PASSWORD=$DB_PASSWORD" \ "GF_DATABASE_SSL_MODE=require" \ "FORCE_SSL=true"
📘 There are many more configuration options available in Grafana. Review Grafana's configuration documentation for more information.
Finally, follow the How do I expose my web app on the Internet? tutorial to expose your Grafana app over the internet. Make sure to use the same domain you configured Grafana with (
$YOUR_DOMAIN in the example above)!
Once you've exposed Grafana, you can navigate to
$YOUR_DOMAIN to access Grafana. Connect using the username
admin and the password you configured above (
Connect to an InfluxDB Database
Once logged in to Grafana, you can connect Grafana to an InfluxDB database by creating a new data source. To do so, click the Grafana icon in the top left, then navigate to data sources and click "Add data source".
The following assumes you have provisioned an InfluxDB database. You'll need to interpolate the following values
$INFLUXDB_HOST: The hostname for your InfluxDB database. This is of the form
$INFLUXDB_PORT: The port for your InfluxDB database.
$INFLUXDB_USERNAME: The username for your InfluxDB database. Typically
$INFLUXDB_PASSWORD: The password.
These parameters are represented by the connection URL for your InfluxDB database in the Aptible dashboard and CLI. For example, if your connection URL is
https://foo:firstname.lastname@example.org:456, then the parameters are:
Once you have those parameters in Grafana, use the following configuration for your data source:
- Name: Any name of your choosing. This will be used to reference this data source in the Grafana web interface.
- Type: InfluxDB
- HTTP settings:
- HTTP Auth: Leave everything unchecked
- Skip TLS Verification: Do not select
- InfluxDB Details: - Database: If you provisioned this InfluxDB database on Aptible and/or are using it for an InfluxDB database metric drain, set this to
db. Otherwise, use the database of your choice. - User:
Finally, save your changes.
Here are a few suggested queries to get started with an InfluxDB metric drain. These queries are designed with Grafana in mind. To copy those queries into Grafana, use the raw text editor mode in Grafana.
📘 In the queries below,
$timeFilterwill automatically be interpolated by Grafana. Leave those parameters as-is.
RSS Memory Utilization across all resources
SELECT MAX("memory_rss_mb") AS rss_mb FROM "metrics" WHERE $timeFilter GROUP BY time($__interval), "app", "database", "service", "host" fill(null)
CPU Utilization for a single App
In the example below, replace
ENVIRONMENT with the handle for your environment and
HANDLE with the handle for your app
SELECT MEAN("milli_cpu_usage") / 1000 AS cpu FROM "metrics" WHERE environment = 'ENVIRONMENT' AND app = 'HANDLE' AND $timeFilter GROUP BY time($__interval), "service", "host" fill(null)
Disk Utilization across all Databases
SELECT LAST(disk_usage_mb) / LAST(disk_limit_mb) AS utilization FROM "metrics" WHERE "database" <> '' AND $timeFilter GROUP BY time($__interval), "database", "service", "host" fill(null)
Once you've added your first data source, you might also want to consider following Grafana's getting started documentation to familiarize yourself with Grafana.
📘 If you get an error connecting, use the
aptible logscommands to troubleshoot.
That said, an error logging in is very likely due to not properly creating the
sessionsdatabase and the
sessiontable in it as indicated in Configuring the database.