Debugging measurements

Debugging containers #

The GMT offers two approaches to debugging that can be used separately or in conjunction

–dev flags #

The first approach is to use the --dev-* switches as defined in the runner switches →.

Here you can turn on many switches that speed up a run. Please note that no useful measurement will come out of the tool. These flags should only be used when debugging.

Especially the --dev-flow-timetravel is extremely useful as it will let you retry a step in the usage scenario without having to go through all previous steps. It will also keep the container state so that if you can do live edits to the files in your local filesystem that are mounted then writeable into the container. Please note that this only works with a local repository. If your repository is online only atm clone it first to your local filesystem. This allows for editing files while running a usage scenario

A typical call looks like this: python3 --uri MY_LOCAL_PATH --name Testing --allow-unsafe --dev-no-metrics --dev-no-sleeps --dev-cache-build --dev-flow-timetravel

–debug flag #

The second approach in debugging a usage_scenario is to
turn the --debug flag of the runner.py on.

When you call the runner.py locally it will turn into
a steppable mode where you continue to the next step by pressing enter.

You can then enter one of the containers to see if
the required services are running correctly.

An example call would be:

docker exec -it MY_CONTAINER_NAME bash

Some container do not have bash. However sh, which has less capabilities,
should be available in most cases.

–print-logs #

If your container fails to boot in the GMT many of the prior mentioned debugging flags do not help.

Since this is usually related to a problem in the Dockerfile you can show the output of the docker build client by adding --print-logs

Debugging containers via HTTP / exposed ports #

If entering the container looks fine and you need to access them through some of their
exposed ports (ex. via Browser through HTTP) turn on the --allow-unsafe flag to bind
the ports specified in the usage_scenario.yml.

Debugging metric providers #

To see if the Metric Providers → are working correctly you have two options:

• Start them manually from their respective folder under /metric-providers/... and look if the output is as expected
• Turn on the --no-file-cleanup switch to see if the files generated in /tmp/green-metrics-tool/[...].log are in expected format