k8s-janitor Service

Service managing k8s cluster with our business logics.

Development

Prerequisites

• .NET Core 2.2 SDK (download)
• Docker (any relatively new version will do)
• All scripts are written in bash. You can use gitbash on windows

Other than the above prerequisites this application requires no additional special setup on your machine. Open the solution or root folder in your editor of choice and start cracking!

Access to Kubernetes

The The application operates on a Kubernetes cluster. The credentials from current-context in ./kube/config will be used if the environment variables KUBERNETES_SERVICE_HOST & KUBERNETES_SERVICE_PORT are not set.

Running the application locally

The folder local-development contains bash scripts that enables you to run the application locally while developing.

• start-dependencies.sh starts a kafka cluster
• watch-run.sh starts the api project with environment variables set for local development and will rebuild on file change
• watch-run-unit-tests.sh runs unit tests and will rerun them on file change

The file rest-request.http contains some rest request you can use for development, you can use the vsts plugin humao.rest-client to execute the requests

Environment variables

The application requires the following environment variables when running locally:

Name	Description
KUBERNETES_SERVICE_KAFKA_BOOTSTRAP_SERVERS	A list of host/port pairs to use for establishing the initial connection to the Kafka cluster.
KUBERNETES_SERVICE_KAFKA_GROUP_ID	Id of the consumer group that the application will join. Only a single consumer in a group will read a message.
KUBERNETES_SERVICE_KAFKA_ENABLE_AUTO_COMMIT	commit the Offset on Consumer fetches or manually.

Running in a container

To run the application you first need to execute a script located in the repository root. This will restore any dependencies and build both the application and also a container image using Docker. Run the following on your command line in the repository root:

./pipeline.sh

Now you should be able to start a container by running the following on your command line:

docker run -it --rm -p 8080:80 k8s-janitor

You should be able to navigate to http://localhost:8080 in a browser.

Running in production

Environment variables

The application requires the following environment variables when running in production:

Name	Description
AWS_S3_BUCKET_REGION	The region the buckets exist in
AWS_S3_BUCKET_NAME_CONFIG_MAP	name of the bucket containing the config map file
CONFIG_MAP_FILE_NAME	Name of the file the configmap will be stored in.
KUBERNETES_SERVICE_KAFKA_BOOTSTRAP_SERVERS	A list of host/port pairs to use for establishing the initial connection to the Kafka cluster.
KUBERNETES_SERVICE_KAFKA_GROUP_ID	Id of the consumer group that the application will join. Only a single consumer in a group will read a message.
KUBERNETES_SERVICE_KAFKA_SASL_PASSWORD	Kafka Simple Authentication and Security Layer password
KUBERNETES_SERVICE_KAFKA_SASL_USERNAME	Kafka Simple Authentication and Security Layer username

Deployment prerequisites

With the change to scoped Kubernetes service connections during deploment, certain manifests have been moved out of the k8s directory and moved to the k8s_initial directory.

The manifests within k8s_initial will have to be run manually or with a different service connection due to elevated rights.

---

If the scoped service account is missing for deployment, see https://wiki.dfds.cloud/en/teams/devex/selfservice/Kubernetes-selfservice-deployment-setup