Quick start

From namespace access to publishing an application via GitOps and the GUI.

This guide walks through application delivery in a namespace: get access, prepare the image and manifests, sync state from Git with Argo CD, publish the application, and verify that it works.

1. Get access to the namespace #

Sign in to the platform GUI, select a cluster, and confirm that the namespace you need is available. To work from the command line, use the procedures in kubectl-shturval-plugin. For API access, see the API description.

On the namespace dashboard you can see workloads, services, ingresses, metrics, quotas, events, and the link to Argo CD if the shturval-cd module is installed and a project has been created for the namespace.

2. Prepare images and manifests #

Before delivering the application, build a container image, publish it to a registry accessible to the cluster, and specify a stable tag in the manifests. Requirements for the image and imagePullSecrets are described in How to deploy an application.

For production workloads, keep manifests in Git, set requests and limits, configure health checks, and avoid the latest tag.

3. Set up delivery via Git and Argo CD #

The platform provides application delivery via GitOps: Argo CD synchronizes cluster state with a Git repository of manifests or Helm charts. How the platform connects Argo CD clusters, projects, and access rights is described in Argo CD.

To create or verify a workload manually in the GUI, see How to deploy an application. It covers Deployment, StatefulSet, DaemonSet, Job, CronJob, and Pod.

4. Publish the application #

After a successful deployment, expose the application with a Service, and configure Ingress for HTTP/HTTPS access from outside the cluster. Service type selection and routing checks are described in Services in a namespace and Ingress in a namespace. For TLS in the namespace, use Issuer.

If the application needs to scale, configure the number of replicas or HPA in Deployment.

5. Check status and diagnose failures #

For day-to-day checks, use the namespace dashboard. It shows workload status, connections to Service and Ingress, metrics, quotas, and events. When something fails, open the pod or workload page to read container logs, review events, or use the interactive pod terminal for manual diagnostics.

The order for checking logs, events, and common errors such as CrashLoopBackOff, ImagePullBackOff, or Ingress issues is described in Pod.