Sparrow is a tool for automating Helm Pull Request Management and Deployment via PR Comments

Helm Pull Request Automation

A Flask application that listens for Helm Chart pull request events via webhooks.

When a PR is opened to a charts Repo tracked by Sparrow, Sparrow will generate the diffs for changes that would occur if the modified charts were applied and add them in the PR comments. Sparrow supports a single chart being configured for different target environments and will generate the diffs against each environement. You may then command Sparrow to apply the changes for all or a specified set of charts and environments via PR comments.

• Make Helm Chart changes visible to your whole team.
• Enable non-operations engineers to collaborate on Helm Chart Configurations.
• Standardize your Helm workflows.

• sparrow diff: Generate diffs for charts changed in the PR
• sparrow apply: Upgrade all of the deployed charts with changes from the PR
  • -f <chart>@<env>: target a subset of charts and envrionments when upgrading by passing one or more targets via the -f flag

In order to use sparrow you must define a sparrowfile.yaml at the root of you project. A SparrowFile provides Sparrow with the configruations for authenticating to clusters, and generating chart diffs for different environments. In addition to the sparrowfile you may optionally add a namespace key and releaseName key to the Charts.yaml for all charts in your repo that specifies desired namespace and Helm release name for your chart. If not specified the Chart name will be used as the release name and the default namespace will be used for releases

Here is an example Sparrowfile:

clusters:
  - name: dev-cluster
    provider: azure
    providerConfig:
      clusterName: aks-dev-cluster
      resourceGroup: aks-dev-cluster-rg
  - name: test-cluster
    provider: azure
    providerConfig:
      clusterName: aks-test-cluster
      resourceGroup: aks-test-cluster-rg

chartConfigurations:
  - path: clusters/teamA/ ## This will configure all charts under the path
    environments: ## Map values files to release environments
      - name: dev
        valuesFiles: 
        - values-dev.yaml
        cluster: dev-cluster
      - name: test
        valuesFiles: 
        - values-test.yaml
        cluster: test-cluster

The following table provides the currently supported Sparrowfile configurations

Sparrow is mildly opionated on the structuring of your charts repo. The following layout is a suggestion for those getting started however the tool is flexible if your project structure is different:

├── README.md
└── clusters
    ├── team1
    │   └── signoz
    │       ├── Chart.lock
    │       ├── Chart.yaml
    │       ├── README.md
    │       ├── charts/
    │       ├── templates/
    │       ├── values-test.yaml/
    │       └── values-dev.yaml
    └── team2
        └── redis
            ├── Chart.lock
            ├── Chart.yaml
            ├── README.md
            ├── charts/
            ├── templates/
            ├── values-dev.yaml
            └── values-test.yaml

An official image is available through ghcr. The creation of a Helm chart is currently in progress. The application can be configured via the following environment variables.

Sparrow can be optionally configured to authenticate to Azure Kubernetes Service. Sparrow supports authentication via MSI if you choose to deploy on a compute instance in Azure. However, you can also configure the application to authenticate via a Service Principal (SP).

• Could turn this into a distributed design by passing receivers events into message broker for processing.
• Support for:
  • Fresh install diffs
  • VCS Platforms:
    • [] Github
  • K8s Providers:
    • [] Local
    • [] GKE
    • [] EKS