Skip to content

Operator onboarding

This guide is for platform operators generating Application Hub configuration.

1. Install and verify

Use one of:

1
2
pip install app-hub-configurator
dump-config --help

or from source:

1
hatch run dump-config --help

2. Generate your first config

1
2
3
4
dump-config \
  --profiles coder_app,gpu_coder_app \
  --groups group-a,group-b \
  --output config.yaml

3. Inspect what is available

1
2
dump-config --list-profiles
dump-config --describe-profile coder_app

4. Apply runtime overrides

1
2
3
4
5
6
dump-config \
  --profiles coder_app \
  --groups group-a \
  --override coder_app:cpu_limit=4 \
  --override coder_app:mem_limit=8G \
  --output config.yaml

5. Use external profile plugins

1
2
3
4
5
dump-config \
  --profiles-dir ./my-profiles \
  --profiles my_custom_profile \
  --groups group-a \
  --output config.yaml
  • docs/getting-started.md
  • docs/cli/overview.md
  • docs/cli/plugins.md
  • docs/cli/troubleshooting.md

7. Skaffold workflow (cluster deployment path)

If you deploy Application Hub through Helm/Skaffold, use the repository skaffold.yaml.

Typical flow:

  1. Generate config.yaml with dump-config.
  2. Deploy using one of the Skaffold profiles:
  3. baseline
  4. develop
  5. main

Example:

1
skaffold run -p develop

Notes:

  • skaffold.yaml injects config.yaml into the chart values (setFiles.configYml).
  • Image/tag selection is profile-based in Skaffold.