Probation Integration Services

A collection of small, domain-focused integrations to support HMPPS Digital services that need to interact with probation data. Typically, these integration services will perform translations between HMPPS and Delius domain concepts, and are responsible for:

publishing REST endpoints to read existing data from the Delius database
listening for HMPPS Domain Event messages and writing data into the Delius database

Project

Goals

This project is intended to reduce the surface area of larger integration systems (e.g. Community API), and to replace other components (e.g. Case Notes to Probation) with simpler services that have direct access to the Delius database.

With this in mind, we aim to:

Support HMPPS Digital teams by building and deploying at pace
Separate overlapping domain concepts by creating smaller, more focused services
Simplify the developer experience by unifying common approaches and streamlining workflows

Tooling

Code is written in Kotlin, using Spring Boot
Built and tested as a multi-project Gradle build
Unit tests with JUnit 5 and Mockito
Integration tests with Spring Boot Test, Wiremock and H2
End-to-end testing with Playwright - see End-to-end tests
Container images are built with Jib, and pushed to GitHub Packages
Code formatting by IntelliJ IDEA formatter, according to Kotlin Coding Conventions
Continuous integration with GitHub Actions

Structure

This repository is structured as a monorepo containing an individually deployable project per integration service.

The directory layout is as follows:

├── .github           ~ GitHub actions workflows and configuration
├── .idea             ~ IntelliJ IDEA configuration
├── buildSrc          ~ Gradle build scripts
├── doc               ~ Technical documentation and decision records
├── libs              ~ Shared Kotlin libraries
├── projects          ~ Source code for integration service projects
│   ├── project-1
│   ├── ...
│   └── project-n
├── templates         ~ Project templates and shared configuration
└── tools             ~ Operational tools e.g. feature flags and database reports

Development

The project is configured to enable developers to build/test/run integration services in isolation without the need for Docker or remote dependencies.

To set up your development environment:

Open the project in IntelliJ IDEA. Select "Import project from external model", then "Gradle".
To run the tests for a service, right-click the src/test or src/integrationTest folder in the project view and select "Run tests" (See Test).
To start the service, use the pre-defined run configuration in .idea/runConfigurations (See Run).

Code formatting

Kotlin code is formatted using IntelliJ IDEA's code formatter, which follows the Kotlin Coding Conventions.

GitHub Actions will automatically fix any formatting issues when you open a pull request. You can also use ⌘⌥L (macOS), or Ctrl+Alt+L (Windows/Linux) to manually reformat your code in IntelliJ IDEA. See Reformat code.

Note: The code formatter does not remove unused imports by default. You should enable Optimise on save in your IntelliJ IDEA settings to ensure you do not commit unused imports.

Build

IntelliJ will automatically build your code as needed. To build using Gradle, follow the instructions below.

Gradle

Any tasks you run from the root project, without specifying a project name will be run on all the children. To build the entire repository using Gradle, run:

./gradlew build

To build just a specific project.

./gradlew <project-name>:build

# For example,
./gradlew tier-to-delius:build
./gradlew workforce-allocations-to-delius:build

Use buildDependents to build and test all projects that depend on a given project (for instance a shared library)

./gradlew <project-name>:buildDependents

# For example,
./gradlew libs:audit:buildDependents

Docker

To build Docker images locally, run:

./gradlew jibDockerBuild

Run

IntelliJ

In IntelliJ IDEA, a Spring Boot run configuration is automatically made available for each service. Select it from the toolbar, and click either Run or Debug. The service will start in the dev profile, which configures any embedded test data and services.

Run configuration files are stored in .idea/runConfigurations.

Gradle

To run Gradle tasks in a subproject, prepend the task name with the name of the project. Environment variables can be used to set the Spring profile. For example,

SPRING_PROFILES_ACTIVE=dev ./gradlew <project-name>:bootRun

Using a live-like database

A Docker image is available for the Delius database, which is pre-populated with all the Delius stored procedures and reference data. To access and run the image, follow the instructions in the HMPPS Delius Docker Images repository.

Once the database is running, use the delius-db profile to connect to it:

SPRING_PROFILES_ACTIVE=dev,delius-db ./gradlew <project-name>:bootRun

Test

Integration tests

Integration tests use WireMock JSON files to mock any external services.

The json files for simulations must reside in simulations/mappings in the dev class path. This makes them usable for dev and test. Any json bodies to add to mappings must live in the simulations/__files directory. These are the defaults for WireMock. Any json mappings and body files provided in these locations will be automatically loaded and available during dev and test.

The WireMock Server is exposed as a spring bean and can be injected into Spring Boot (Integration) Tests for verification or adding extra scenarios specific to a test that are not available in json.

@Autowired
private val wireMockServer: WireMockServer

The strategy with the dev/test profiles is to use a single WireMock server and distinguish any potential duplicate urls using a service name on the url if required. For example if two urls were used as part of a service

https://hmpps.service1/offender/{crn}
https://hmpps.service2/offender/{crn}

When mocking these urls the following would be appropriate (rather than a separate mock server for each service)

{wiremockUrl}:{wiremockPort}/service1/offender/{crn}
{wiremockUrl}:{wiremockPort}/service2/offender/{crn}

All other concepts of Spring Boot Tests are usable as per Spring documentation.

Testing with an Oracle Database

By default, integration tests run against an in-memory H2 database. This can sometimes cause problems where H2 doesn't behave in the same way as the target Oracle database.

To run the integration tests using a slim Oracle database container, use the oracle profile:

SPRING_PROFILES_ACTIVE=integration-test,oracle ./gradlew integrationTest

End-to-end tests

End-to-end tests are written in TypeScript using Playwright, in the hmpps-probation-integration-e2e-tests repository.

We run the end-to-end tests in GitHub Actions as part of the deployment pipeline, against a real dev/test environment with all dependencies. Running the tests in a real environment gives us confidence that all the integration points involved in the user journey are working correctly before we push changes to production.

You can also run the end-to-end tests from your branch to get early feedback, by following the instructions here: Manually running a workflow. Note: this will deploy your dev code to the test and preprod environments.

To access internal services from GitHub Actions, we use a repository-level self-hosted runner in MOJ Cloud Platform. See 06-github-actions-runner.yaml. For more information on how this is implemented, see PI-340.

Feature flags

This project uses Flipt to control the availability of certain features. Feature flags allow us to turn on or off parts of a service in production, decoupling "releases" from "deployments".

Feature flags are managed in the Flipt dashboard. You'll need to be in the ministryofjustice organisation to access it.

To add a feature flag to your code:

Create a new boolean flag in the dev, preprod, and prod dashboards.
Update your code to inject the FeatureFlags service, and call enabled("<key>"). Example:

@Service
class MyService(private val featureFlags: FeatureFlags) {
    fun myMethod() {
        if (featureFlags.enabled("my-flag")) {
            // Feature is enabled, do something
        } else {
            // Feature is disabled, do something else
        }
    }
}

For more information about Flipt, check out the documentation.

Deployment

Once the code is built and tested, GitHub Actions deploys the updated images for each service to MOJ Cloud Platform. An AWS Transit Gateway connection enables secure access to dependencies in the Delius account, such as the Delius database.

Environments

Although the services are deployed to MOJ Cloud Platform namespaces, they typically need to interact with resources in the Delius environments.

We map MOJ Cloud Platform namespaces to Delius environments as follows:

MOJ Cloud Platform	Delius	Used for
hmpps-probation-integration-services-dev	delius-test	End-to-end testing
hmpps-probation-integration-services-preprod	delius-pre-prod	Testing with live data
hmpps-probation-integration-services-prod	delius-prod	Live service

Configuration

Each subproject has a deploy folder containing YAML files used for configuration. The standard values.yaml file provides common configuration across all environments, while additional files (e.g. values-dev.yml) can be used to set environment-specific configuration.

There is also a repository-level helm-defaults.yml containing the default configuration across all projects.

├── templates
│   └── helm-defaults.yml                 # default values across all projects
└── projects
    └── workforce-allocations-to-delius
        └── deploy
            ├── values.yaml               # common values across each environment
            └── values-<environment>.yml  # 1 per environment

Setting project wide values

<project>/deploy/values.yaml

📝 the file extension for this file must be .yaml, not .yml, due to restrictions in Helm.

This file contains values that are the same across all environments. Example:

generic-service:
  # Image
  nameOverride: project-name
  image:
    repository: ghcr.io/ministryofjustice/hmpps-probation-integration-services/project-name
    port: 1234

  # Container resources
  replicaCount: 2
  resources:
    limits:
      cpu: 1
      memory: 1G

  # Environment variables
  env:
    JAVA_OPTS: "-Xmx512m"
    SERVER_PORT: "8080"
    SPRING_PROFILES_ACTIVE: "my-profile"

  # Secrets
  # These are stored in the Kubernetes namespaces
  namespace_secrets:
    common:
      SPRING_DATASOURCE_URL: DB_URL
    project-name:
      CLIENT_SECRET: CLIENT_SECRET

# Slack alerts
generic-prometheus-alerts:
  targetApplication: project-name

Setting environment specific values

<project>/deploy/values-<environment>.yml

This file should only contain values that differ between environments.

Example:

generic-service:
  resources:
    limits:
      memory: 2G

  env:
    SERVICE_URL: https://example.com

Accessing the Delius Database

To configure access to the Delius probation database, add an access.yml file to the project's deploy/database folder.

The access.yml file defines the account used for accessing the database, as well as an optional user for auditing interactions. Example (see access.yml:

database:
  access:
    username_key: /prison-case-notes-to-probation/db-username    # references AWS Parameter Store 
    password_key: /prison-case-notes-to-probation/db-password    # (see Secrets section above)
    tables:
      # A list of tables that the service can write to. Read access is granted on all tables.
      - audited_interaction
      - contact

  audit:
    username: PrisonCaseNotesToProbation
    forename: Prison Case Notes
    surname: Service

When this file is merged to main, the Database Access workflow in GitHub Actions will invoke a Systems Manager Automation Runbook in AWS to create/update the access account and the audit user in the Delius database. The runbook is in the hmpps-delius-pipelines repository.

Support

For any issues or questions, please contact the Probation Integration team via the #probation-integration-tech Slack channel. Or feel free to create a new issue in this repository.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

README.md

README.md

Probation Integration Services

Project

Goals

Tooling

Structure

Development

Code formatting

Build

Gradle

Docker

Run

IntelliJ

Gradle

Using a live-like database

Test

Integration tests

Testing with an Oracle Database

End-to-end tests

Feature flags

Deployment

Environments

Configuration

Setting project wide values

Setting environment specific values

Accessing the Delius Database

Support

Files

README.md

Latest commit

History

README.md

File metadata and controls

Probation Integration Services

Project

Goals

Tooling

Structure

Development

Code formatting

Build

Gradle

Docker

Run

IntelliJ

Gradle

Using a live-like database

Test

Integration tests

Testing with an Oracle Database

End-to-end tests

Feature flags

Deployment

Environments

Configuration

Setting project wide values

Setting environment specific values

Accessing the Delius Database

Support