We have recently rebranded and changed our name from Dataloop.IO to Outlyer. Our agent is still called
dataloop agent, and relevant code reflects the old name (Dataloop) as well. Thank you for your patience as we update everything.
The Outlyer Agent can run plugins that scrape Prometheus http endpoints. We recommend that you install a Outlyer Agent on each server and then create a separate plugin to monitor each endpoint on localhost.
Example for Node Exporter¶
Start the Prometheus Node Exporter on a server following the instructions here:
Then create a new plugin in Outlyer called
Paste in the following code to scrape the endpoint:
1 2 3
#!/usr/bin/env python import requests requests.get('http://localhost:9100/metrics').text
Where port 9100 is the port the Node Exporter is running. Select the Outlyer Agent and press run to confirm that metrics are being received.
Now set the output format to Prometheus and specify the scrape interval (the default is to scrape every 30 seconds) on the plugin details page. Save and apply to either a single Outlyer Agent or a tag for deployment.
A list of other exporters can be found here:
Example for Application Instrumentation¶
Select a Prometheus client library for your language.
Each library has a set of documentation for how to instrument your code. In our example we'll pick a Python app for payments processing. The example shows how to import the client library and expose metrics on http://localhost:8000.
Create a plugin called
payments.py with the following content:
1 2 3
#!/usr/bin/env python import requests requests.get('http://localhost:8000/metrics').text
Test this by pressing the run button against a node running the payments application. Remember to switch the plugin format to Prometheus on the plugin details page. Save and apply to either a single Outlyer Agent or a tag for deployment.