Skip to main content

🚀 Site Values Auto Deployment

This guide will show you how to structure your configurations and quickly troubleshoot the site-values-deployment pipeline like a pro!

🔄 Standard Deployment Workflow​

Triggering a deployment is effortless: just push your configuration changes straight to the master branch.

Example: Need to update Loki in the dev environment?

  1. Open up infra/loki/dev.yaml and make your edits.
  2. Commit and push your changes to master.
  3. The pipeline instantly detects the change, maps it to the infra-loki release, and deploys only that specific release to dev. Magic! ✨

📂 File and Directory Structure​

To guarantee the pipeline detects your changes, you must stick to the expected Helmfile directory structure. Don't worry, it supports nested Helmfiles!

A standard, clean service layout looks exactly like this:

team-folder/
├── helmfile.yaml.gotmpl          # Defines the releases
├── service-name/
│   ├── base.yaml.gotmpl          # Base config (applies everywhere!)
│   ├── dev.yaml                  # Dev-specific config
│   ├── prod.yaml                 # Prod-specific config
│   └── dev/                      # Additional dev-specific overrides
│       └── additional-config.yaml

🧪 Running a Dry Run​

Want to test the waters before diving in? You can preview what the pipeline would do without actually applying anything to the cluster!

Simply run the pipeline manually in Jenkins and flip the DRY_RUN parameter to true. The pipeline will run a safe helmfile diff, print the results for you to inspect, and skip the final apply stage.

🛠️ Troubleshooting​

Did your deployment skip a beat? Check these common culprits:

  • 👻 No changes were detected: Double-check that your modified file is inside a monitored team folder and is explicitly listed in the values section of your release inside helmfile.yaml.gotmpl. Also, ensure the filename perfectly matches the target environment!
  • 💥 Too many releases deployed: Oops! Check your file naming. You might have accidentally dropped environment-specific configs into a base file, triggering a massive global deployment.
  • 🛑 A dynamic stage failed: If the apply process fails midway, it will not auto-retry. Manually trigger it in Jenkins with the DEBUG_LOGGING=true parameter enabled. This will spit out detailed trace logs to help you hunt down the root cause!