How to configure Application Performance Monitoring

This example uses Elastic APM in a Django project. However, the principles are the same for other projects and services.

Set up the APM instance

Log in via the Elastic control panel.

Create a new Elastic deployment. It makes sense to select the same cloud platform and region you’re using on Divio - select AWS US East unless you know otherwise, and the default I/O Optimized configuration.

Note the username and password provided for future reference. The APM instance created will also have a secret token; make a note of this too. Finally, you will need to copy the APM Server URL. (It takes a few moments to create and the instance and for all this information to become available.)

Other services will allow you to create an instance and will provide the credentials for using it in much the same way.

Install and configure the Divio Telemetry addon

Divio Telemetry is available as an Aldryn addon via the Control Panel. It uses the official elastic-apm package provided by Elastic, and adds easy configuration via a single environment variable, DEFAULT_APM_DSN.


At the time of writing, you will need to pin the elastic-apm package in your


as the result of a known issue.

DEFAULT_APM_DSN is a URL, and can be assembled from the Elastic credentials you collected earlier. It’s in the form:

https://<Divio project slug>:<secret token>@<APM Server URL>:443

Your own URL should look something like:

It should not include the protocol fragment (https://) of the APM Server URL.

Apply this as an environment variable, either using the Control Panel, or in the .env-local file. It makes sense to use a different APM instance for each environment.

Using a different provider, you will need to install the appropriate software in your project and provide it with the necessary credentials to access the third-party service.

Check the results

When the project has been rebuilt/deployed as required, you can visit the APM Server URL. Log in with the username and password. Your new instance will be displayed, along with the data it has collected.