Deploy the Edge Delta Agent with a Profiler

George Alpizar
George Alpizar
  • Updated


You can use this document to learn how to deploy the Edge Delta Agent with a profiler.

The Edge Delta Agent is a Go-based binary that (when specified during deployment) can expose CPU and memory profiling information.

This information can be useful to understand specific activity, such as understanding why CPU usage is high.

After deployment, to obtain the data, you can run a command to download a report. 

Deploy the Agent with a Profiler


You can only add a profiler while you deploy an agent.

You cannot add a profiler to an existing agent deployment.  

To deploy the agent with a profiler, follow the regular installation instructions for each agent type; however, you must add PROFILER_PORT for port 6060 to the installation command.

Review the following examples.


In the examples below, replace the ED API KEY with your own key.

Linux Example

For instance, for Linux, the regular deployment command is: 

To deploy with a profiler, the command would be:

Docker Example

In another example, for Docker, the regular deployment is:

  • docker run --rm -d --name edgedelta -v /var/run/docker.sock:/var/run/docker.sock:ro -e "ED_API_KEY=f1111e-e1d1-1ad1-b11d-d1a11111b1"

To deploy with a profiler, the command would be: 

  • docker run --rm -d --name edgedelta -v /var/run/docker.sock:/var/run/docker.sock:ro -e "ED_API_KEY=f1111e-e1d1-1ad1-b11d-d1a11111b1" -e "PROFILER_PORT=6060"

Review the following installation documents. Remember to add PROFILER_PORT for port 6060 to the installation command:

Obtain a Profiler Report 

When you obtain a profiler report, you can send the report to Edge Delta for analysis.


You can also view the data using the pprof tool.

  1. Run the following command to install pprof, which requires golang installed: go install
  2. Run the following command to visualize pprof results: pprof -web /tmp/cpu.pb.gz

Based on the information you want to obtain, run the following command: 

For information on heap, run: 

For information on CPU, run:

For information on goroutine, run: 

Share this document