Table of Contents:

• Core Concepts
• Workload Management
• valohai.yaml - Config File
  • step
  • step.name
  • step.image
  • step.command
  • step.parameters
  • step.inputs
  • step.environment
  • step.environment-variables
  • Special environment variables
  • pipeline
  • pipeline.nodes
  • pipeline.edges
  • endpoint
  • endpoint from a command
  • endpoint from a WSGI definition
  • Examples
• valohai-cli - Command-line Client
• Jupyhai - Jupyter Notebooks with Valohai
• Valohai API for Automation
• Docker Images
• On-premises Installations
• Architecture Overview
• Frequently Asked Questions
• Tutorials

/ valohai.yaml - Config File

step.environment

Environment defines the hardware and the location where to run your code.

Environments encapsulate the following information:

• Who owns the environment instances?
• Are you running on Azure, Amazon, Google, OpenStack or on-premises?
• Are you running in USA, Asia, Europe, etc.
• How much memory and CPU is available?
• Does it have GPUs?

Environment defined in the YAML is the default value for executions of that particular step type. In other words, it can overwritten by selecting a separate environment in the web interface dropdown or with vh exec run -e <SLUG> <STEP_NAME> command-line parameter.

You can use vh environments command to see available environments after logging in.

The most common environment slug syntax is <OWNER_IF_NOT_VALOHAI>-<CLOUD_PROVIDER>-<LOCATION>-<HARDWARE_TYPE>. For example:

• environment: aws-eu-west-1-g3s-xlarge
• environment: acme-azure-westeurope-nc24
• environment: awesome-corporation-gcp-europe-west4-b-custom-2-8192

Tip

It is wise to use an environment in the same data center or geographical location as your data store to minimize the data transfer durations.

Was this useful?

💩 🤷 👍

Thank you!