Service Deployment

1. System Context

Developers use Nitric to create services or functions within their application.

Application code is written in files that matches the pattern(s) in the nitric.yaml config file.
The Nitric CLI builds container images for their Lambda functions and push them to a container registry.

Operations use default or overridden IaC (e.g Terraform modules) to provision the necessary resources for their target cloud.

Example AWS Provider

AWS ECR (Elastic Container Registry) stores container images for Lambda functions.
AWS Lambda runs serverless functions based on the container images.
AWS IAM manages roles and policies for secure access to AWS resources.
Docker or Podman is used to build and tag container images before pushing them to ECR.

Example GCP Provider

Docker builds and tags the image, which is then pushed to Google Artifact Registry (GCR).
Google IAM ensures secure access, with the appropriate permissions for the Cloud Run service and service accounts.
The Cloud Run service will run based on the container image pulled from GCR.

2. Container

General-purpose workers that handle tasks like processing queues, topics, or schedules. Services abstract the runtime’s ability to route tasks and events to application-defined logic.

API & Event-Driven Communication

Data Storage & Management

Task Execution & Scheduling

3. Component

Service Module

Configures Terraform to handle the deployment and management of containerized services, abstracting away provider-specific details.
Dynamically creates and manages a container registry for storing service container images, ensuring secure and efficient access.
Automates authentication and tagging for container image pushes, supporting seamless integration with deployment pipelines.
Creates a role with least privilege permissions for executing the service, including necessary trust relationships and policies for interacting with other resources.
Configures containerized services with runtime parameters like environment variables, memory limits, and execution timeouts to optimize performance and scalability.
Optionally supports advanced networking configurations like VPC settings for secure and isolated deployments.
Abstracts the underlying infrastructure for running serverless or containerized services, enabling developers to focus on application logic while providing a consistent interface for operations teams.

4. Code

Developers write application code that implements handlers for the api, bucket, websocket, topic, schedule resources. This code is written in files that matches the pattern(s) in the nitric.yaml file.

Nitric service configuration - nitric.yaml

name: service-name
services:
    - match: services/*.ts
      runtime: ""
      type: ""
      start: npm run dev:services $SERVICE_PATH

HTTP Route Handler

import { api } from '@nitric/sdk'

const customerRoute = api('public').route(`/customers`)

customerRoute.get((ctx) => {
  // construct response for the GET: /customers request...
  const responseBody = {}
  ctx.res.json(responseBody)
})

Bucket On Read/Write/Delete Handler

import { bucket } from '@nitric/sdk'

const assets = bucket('assets')

const accessibleAssets = bucket('assets').allow('delete')

// The request will contain the name of the file `key` and the type of event `type`
assets.on('delete', '*', (ctx) => {
  console.log(`A file named ${ctx.req.key} was deleted`)
})

Operations will use or extend the Nitric Terraform reference modules: