arvados-cluster-activity 3.2.1

Running as a workflow from Workbench

We provide a CWL workflow to generate this report. It’s available as a single file in the Arvados source and included with this Python package. You can register the workflow on your cluster by running:

arvados-cwl-runner [--project-uuid=UUID] --create-workflow cluster-activity.cwl

Then you can launch the workflow from Workbench. All inputs have documented formats and values.

Running as a workflow from the command line

Alternatively, you can run the workflow directly with arvados-cwl-runner. Write an input file following this YAML template:

# Report start date as a `YYYY-MM-DD` string
reporting_start: "YYYY-MM-DD"

# Report end date as a `YYYY-MM-DD` string. Default today.
#reporting_end: "YYYY-MM-DD"

# The base URL of your Arvados cluster's Prometheus server, like
# `https://prometheus.arvados.example/`
#prometheus_host: ""

# Prometheus API token
#prometheus_apikey: ""

# Prometheus API username
#prometheus_user: ""

# Prometheus API password
#prometheus_password: ""

# A string with a Python regular expression.
# Workflows whose name match the expression will be excluded from the report.
#exclude: ""

# A boolean. If true, individual workflow steps will be reported alongside
# their parent workflows.
include_workflow_steps: false

Then run the workflow like this:

arvados-cwl-runner [--project-uuid=UUID] [options ...] cluster-activity.cwl YOUR-INPUTS.yml

Running as a command line tool

This Python package provides a command line tool you can run to generate reports on your own system. Install it with pipx like:

pipx install "arvados-cluster-activity[prometheus]"

If you don’t have a Prometheus server or don’t want Prometheus support, remove [prometheus] from the command line. Advanced users can install the tool to their own virtualenv or elsewhere.

The command line tool provides options to control the report generation. These correspond to the workflow inputs. Run the tool with --help for the full list:

arv-cluster-activity --help

The tool gets Arvados credentials the same as other client tools: it reads the ARVADOS_API_HOST and ARVADOS_API_TOKEN environment variables if those are set, or the ~/.config/arvados/settings.conf file if they are not.

The tool gets Prometheus credentials from the PROMETHEUS_HOST, PROMETHEUS_APIKEY, PROMETHEUS_USER, and PROMETHEUS_PASSWORD environment variables. The values follow the format of the workflow inputs above. You can write these environment variables in a dedicated file and load that with the tool’s --prometheus-auth option.