Orbit Console user guide

Documentation home · Administrator guide

Before you start

Open the console URL supplied by your administrator in a browser with JavaScript enabled. Complete any upstream sign-in required by your organisation. Orbit Console itself does not provide accounts or a login screen.

You need a registered kubeconfig with credentials that permit the intended operations. Requests to Kubernetes run from the console server using those credentials, rather than your browser's identity. Everyone with access to this console can use the registered configurations; only register credentials approved for that audience.

Register and select a cluster

If the administrator has already registered configurations, select one in the Kubeconfig selector at the top of the page, then choose a Context. The file's current context is selected when available; (default) leaves context selection to the kubeconfig loader. A single kubeconfig can contain contexts for several clusters or credentials.

To register a configuration yourself:

1. Click Add in the top bar.
2. Enter a Config name of up to 120 characters that identifies the environment.
3. Paste the complete configuration into Kubeconfig YAML.
4. Click Register, then select the intended configuration and context.

Registration stores the kubeconfig on the console server. Successfully registering a file does not verify connectivity or permission to use its cluster. Its API server address must be reachable from the console server. Certificate files, tokens stored in external files, and authentication helper programs must also be available there; ask your administrator for a suitable configuration.

If the top-bar Add and Delete controls are disabled, the administrator has locked configuration registration and removal with CONTEXTS_READONLY. You can still select existing configurations and contexts, and Kubernetes operations remain available according to the credentials' permissions.

Browse resources

Use the Namespace selector above the resource tree:

ALL does not include cluster-scoped resources. Switch to Cluster to find those objects.

The tree shows only resource kinds with objects that the console can list in the selected scope. Each branch includes its object count, scope and API version. Expand a kind and select an object to see its details. A kind can appear under multiple API versions.

The detail panel shows identity, labels and status information. Workloads such as Deployments, StatefulSets, DaemonSets, Jobs and CronJobs also show related pods. Select a pod from the resource tree to use its Logs or Shell actions.

Resource discovery and the tree are loaded on selection changes, rather than continuously watched. Reload the page to refresh the registry and tree, then reselect your configuration, context and scope. Reselecting an object fetches its current details.

An empty tree can mean that no objects exist, that the selected scope is wrong, or that the credentials cannot list those resources. Listing errors for individual resource kinds are skipped, so an empty tree is not proof that the cluster is empty.

Understand why an object exists

Select an object and click Why? to open Why Exists?. The dialog follows Kubernetes owner references, preferring the controller owner, and displays a chain such as:

Pod web-abc -> ReplicaSet web-xyz -> Deployment web -> OrbitDeployment website

An OrbitDeployment is appended when the console finds a matching entry in its status.managedObjects. The dialog may also show Orbit Deploy metadata hints. Close the dialog with Close.

The chain is an explanation based on available metadata, rather than a complete dependency graph. Missing owners, insufficient permissions, or an unavailable Orbit Deploy API can prevent a complete result. For namespaced objects, Orbit Deploy lookup checks the object's namespace and kube-system; for cluster-scoped objects it lists OrbitDeployments across namespaces.

View pod logs

1. Select a Pod, then click Logs.
2. Choose a Container if the pod contains more than one.
3. Leave Refresh at NO for a single fetch, or choose 1s, 3s, 5s or 10s to poll for updated logs.
4. Click Back to return to the object details and stop polling.

Each fetch returns up to the last 300 log lines; polling replaces the displayed snapshot. This is not a continuous log stream or a view of previous container instances. The container defaults to the kubectl.kubernetes.io/default-container annotation when present, otherwise the first regular container. The selector lists regular containers, not init or ephemeral containers.

Open a pod shell

1. Select a Pod, then click Shell.
2. Choose the desired Container. Changing it opens a new session.
3. Type commands in the terminal. Commands execute inside the selected container with that container's operating-system permissions.
4. Type exit, send Ctrl-D, or click Back to close the session.

The container must provide /bin/sh, and its credentials must allow pod exec. Use Reconnect after a closed session to open another shell. A closed terminal stops accepting input. Opening the shell requires browser access to the terminal assets on cdn.jsdelivr.net and a working WebSocket connection through the console's proxy.

Edit an object

Confirm the configuration, context, namespace and object name before changing a resource. For a resource managed by Orbit Deploy or another controller, update its source configuration when you need a lasting change; reconciliation may overwrite direct edits.

1. Select the object and click Edit.
2. Change values in the generated form. Fields use a CRD schema when available, otherwise the form is inferred from the current object. Array sections provide a + control to append entries.
3. Click Save to submit a merge patch and return to the refreshed details. Use Cancel or Back to leave without saving.

The form excludes status and selected server-managed metadata. It submits the populated form fields, including unchanged values. Empty strings and blank numeric inputs are omitted, so clearing an input does not reliably remove a field or set an empty value. Lists use merge-patch replacement semantics; complex list changes and keys containing dots need particular care. Use your normal Kubernetes tooling for precise field removal, raw YAML, or complex patches.

Kubernetes validates the patch and enforces permissions. A failed save may mean the credentials lack patch access, a value is invalid, or a field is immutable. Inspect the error and reopen the object to check its current state before retrying.

Delete an object or remove a configuration

The Delete button in the object detail toolbar deletes the selected Kubernetes object after confirmation. Verify the kind and namespace/name in the prompt. Kubernetes deletion can be asynchronous while finalizers run, and a controller may recreate an object it manages. There is no console undo action.

The Delete button in the top bar removes the selected kubeconfig registration and its stored file after confirmation. It does not delete Kubernetes objects or revoke the credentials at the cluster. This removal affects everyone using that console instance.

Visit an ingress route

Select an Ingress and use Visit next to a displayed route to open it in a new tab. The console builds links from ingress rules and TLS hosts. Your browser must be able to resolve and reach the target, and the application may require its own authentication. A Visit link does not verify that the route is healthy.

Troubleshooting

When reporting a problem, include the configuration label, context, scope, kind, namespace/name, action and error message. Keep kubeconfig credentials and sensitive log contents out of the report.