@gunzip/opex-dashboard-ts
v0.3.3
Published
Generate operational dashboards from OpenAPI specifications
Maintainers
Readme
OpEx Dashboard
Generate standardized Operational Excellence dashboards from OpenAPI 3 specifications.
Features
- Automatic Dashboard Generation - Parses OpenAPI 3 specs to create standardized operational dashboards
- Multiple Cloud Providers - Azure (current), AWS CloudWatch, Grafana (planned)
- High Performance - Parallel rendering with deterministic ordering
- Type-Safe - Full TypeScript with Zod runtime validation
OpEx Dashboard is distributed as an npm package with two components:
- CLI tool (
opex_dashboard) - Command-line interface for end users - TypeScript library - Programmatic API for integration
Usage
npx @gunzip/opex-dashboard-ts generate --helpLocal Development
git clone https://github.com/gunzip/opex-dashboard-ts.git
cd opex-dashboard-ts
pnpm install
pnpm run build
pnpm run devQuick Start
1. Create Configuration File
# config.yaml
oa3_spec: ./openapi.yaml # or HTTP URL
name: My API Dashboard
location: West Europe
resource_type: app-gateway
data_source: /subscriptions/xxx/resourceGroups/my-rg/providers/Microsoft.Network/applicationGateways/my-gtw
action_groups:
- /subscriptions/xxx/resourceGroups/my-rg/providers/microsoft.insights/actionGroups/my-alerts2. Generate Dashboard
# Output to stdout
npx @gunzip/opex-dashboard-ts generate -t azure-dashboard-raw -c config.yaml
# Save as Terraform package
npx @gunzip/opex-dashboard-ts generate -t azure-dashboard -c config.yaml --package ./output3. Deploy with Terraform
cd output/azure-dashboard
terraform init -backend-config=env/dev/backend.tfvars
terraform apply -var-file=env/dev/terraform.tfvarsDashboard Components
For each endpoint in the OpenAPI spec, the dashboard includes:
Graphs
- Availability: HTTP success rate (status codes < 500)
- Response Codes: Segmentation of all HTTP status codes (1XX, 2XX, 3XX, 4XX, 5XX)
- Response Time: 95th percentile response time
Alarms
- Availability Alarm: Triggers when availability drops below threshold (default: 99%)
- Response Time Alarm: Triggers when response time exceeds threshold (default: 1 second)
Configurable Parameters
For each alarm, you can configure:
- Timespan (Default: 5m) - The aggregation window
- Evaluation Frequency (Default: 10 minutes) - How often to evaluate the rule
- Time Window (Default: 20 minutes) - Data fetch window (must be ≥ evaluation frequency)
- Event Occurrences (Default: 1) - Number of events needed to trigger alert
NOTE: Maximum event occurrences = time window ÷ timespan. For example, with a 30m window and 5m timespan, max is 6 events.
Usage
CLI Commands
opex_dashboard generate [options]
Options:
-t, --template-type <type> Template type: azure-dashboard or azure-dashboard-raw (required)
-c, --config <path> Path to YAML config file, use - for stdin (required)
--package [path] Save as package in directory (default: current dir)
-h, --help Display help
-V, --version Display versionConfiguration
Create a YAML configuration file:
# yaml-language-server: $schema=./config.schema.json
# Required fields
oa3_spec: string # Path or URL to OpenAPI 3 specification
name: string # Dashboard name
location: string # Azure region (e.g., "West Europe")
data_source: string # Azure resource ID
action_groups: string[] # Array of Azure Action Group IDs
# Optional fields (with defaults)
resource_type: app-gateway | api-management # Default: app-gateway
timespan: string # Default: 5m
evaluation_frequency: integer # Default: 10 (minutes)
evaluation_time_window: integer # Default: 20 (minutes)
event_occurrences: integer # Default: 1
availability_threshold`: float # Default: 0.99 (99%)
response_time_threshold: float # Default: 1.0 second
# When generating Terraform packages (using `--package` option),
# you can optionally configure environment-specific settings
terraform:
environments:
dev:
prefix: string # Max 6 chars (required)
env_short: string # Max 1 char: 'd', 'u', 'p' (required)
backend: # Optional backend state configuration
resource_group_name: string
storage_account_name: string
container_name: string
key: string
uat: # Similar to dev
prod: # Similar to devSee examples/ directory for complete configuration samples.
JSON Schema Validation
A JSON Schema is automatically generated from the TypeScript types and included
at config.schema.json. This enables IDE autocomplete
and validation for your configuration files.
To enable schema validation in VS Code and other compatible editors, add this comment at the top of your YAML configuration file:
# yaml-language-server: $schema=https://raw.githubusercontent.com/gunzip/opex-dashboard-ts/refs/heads/main/config.schema.jsonThe schema is:
- Automatically generated during build from Zod schemas using
pnpm run build - Synchronized with TypeScript types - any changes to configuration structure are reflected immediately
- Versioned - matches the package version for compatibility tracking
- Distributed with the npm package for programmatic access
CI/CD Integration
Using the GitHub Workflow from Another Repository
You can use the reusable workflow to automate dashboard generation in your own
repository. Create a .github/workflows/generate-dashboard.yml file in your
repo:
name: Generate Dashboard
on:
push:
branches: [main]
pull_request:
branches: [main]
jobs:
generate:
uses: gunzip/opex-dashboard-ts/.github/workflows/generate.yml@main
with:
# mandatory inputs:
config_path: ./config/dashboard-config.yaml
output_dir: ./generated
# optional inputs with defaults:
pr_title: "Update operational dashboard"
pr_body: "Automated update of dashboard terraform from OpenAPI spec"
base_branch: mainThis workflow will:
- Generate the dashboard using your configuration
- Commit changes to a new branch
- Create a pull request for review
Workflow Inputs
config_path(required): Path to your YAML configuration fileoutput_dir(required): Directory to save generated filestemplate_type(optional): Template type (azure-dashboardorazure-dashboard-raw, default:azure-dashboard)pr_title(optional): Title for the generated pull requestpr_body(optional): Description for the generated pull requestbase_branch(optional): Base branch for the pull request (default:main)
Development
Setup
pnpm install
pnpm run buildScripts
pnpm run build # Bundle with tsup (ESM output)
pnpm run dev # Watch mode for development
pnpm run typecheck # TypeScript compilation check
pnpm test # Run all tests (unit + integration)
pnpm run test:coverage # Coverage report (80% thresholds)
pnpm run test:unit # Unit tests only
pnpm run test:integration # Integration tests only
pnpm run test:watch # TDD watch mode
pnpm run lint # ESLint with auto-fix
pnpm run lint:check # Lint check only
pnpm run format # Prettier format
pnpm run format:check # Format check onlyTesting
The project includes comprehensive unit and integration tests with Vitest:
- Unit Tests (
test/unit/) - Test individual functions and components - Integration Tests (
test/integration/) - Test end-to-end workflows
Run tests with automatic cleanup:
pnpm test # Run all tests
pnpm run test:coverage # Run with coverage report
pnpm run test:watch # Watch mode for TDDTests automatically:
- Clean up temporary directories after each test
- Use unique temp directories to avoid conflicts
- Validate outputs against existing fixtures
Code Quality
pnpm run lint # ESLint with auto-fix
pnpm run lint:check # Lint check only
pnpm run format # Prettier format
pnpm run format:check # Format check only
pnpm run typecheck # TypeScript compilation checkCode Style
- TypeScript strict mode enabled
- No
anytypes - Useunknownwith type guards - Small focused files (< 200 lines)
- Zod validation for runtime safety
- File headers explaining module purpose
- Colocated schemas (
*.schema.ts)
