Docs / Working with resources

Working with resources

Actions, safe YAML edits with server-side apply, applying to several clusters, rollouts, history and bulk actions.

Actions and equivalent commands

Every action dialog (scale, restart, rollout pause, resume and rollback, trigger, cordon, drain, delete, port forward, debug, Node shell, apply, and Helm install, upgrade and rollback) has an Equivalent command section with the kubectl or helm command that does the same thing, ready to copy. Every resource's actions menu also has Copy kubectl get.

Editing YAML

Edit any resource on its YAML tab. Changes are applied with server-side apply, as field manager klystiq, after a dry run shows you the diff:

  1. Edit, then press Cmd/Ctrl+S to review. Klystiq runs a server-side dry run and shows exactly what would change.
  2. Press Cmd/Ctrl+Enter to apply.

Fields owned by other managers, such as a GitOps controller or another tool, produce a conflict you can review and, if you mean it, force.

Secret values are masked until you reveal them, one value at a time.

New resources, in several clusters at once

New resource (the + button) takes a manifest to one or several open contexts:

  • Review runs a dry run in each writable context, and compares with the live object in each read-only one. It then lists every context with what would happen: creates it, updates it, no change, differs, or the error, with a diff.
  • Apply writes to the writable contexts where the review passed and something changes, and reports each result.

The same checks as Best practices run on the manifest before you apply it.

Read-only contexts

Mark a context read-only in its settings and Klystiq refuses every write to it, in the Rust command layer as well as the interface. That covers applies, scaling, restarts, deletes, shells, attach and debug, file uploads, GitOps actions and agent proposals. Reading, logs, comparisons, the policy side of connection tests and permission checks still work.

Read-only mode is Klystiq's own: it doesn't limit commands you type in a local terminal.

Rollouts

After you restart, scale, resume or roll back a Deployment, StatefulSet or DaemonSet, or apply one from YAML or New resource, a card in the bottom corner follows its rollout the way kubectl rollout status does. It shows what the rollout is waiting for (new replicas updated, old ones terminating, updated ones becoming available), with a progress bar and the time so far.

When a new Pod fails in a way waiting won't fix (crash loop, image pull, bad config, unschedulable, out of memory), the card says Stuck. It shows which Pod and why, the last lines of its previous log and recent warning events, with Logs, Ask AI and Roll back… beside them. A Deployment past its progress deadline shows Failed. Watch rollout in the actions menu follows one without changing anything.

History

Every resource has a History tab:

  • Changes seen: the versions Klystiq saw while it ran, each with the fields that changed and a side-by-side diff. Status and server bookkeeping are left out, and Secret values appear only as fingerprints.
  • Rollout revisions: the revisions Kubernetes keeps for Deployments, StatefulSets and DaemonSets, including ones from before Klystiq started, with the Pod template diff between them. Roll back to revision N… shows the diff from the running template before you confirm.
  • Field owners: which tool last changed which fields, and when.

History is kept in memory (30 versions per object) and is gone when the app quits.

Used by

The Overview of a ConfigMap, Secret, ServiceAccount, PersistentVolumeClaim or Service lists what uses it: workloads through their Pod templates, Ingress TLS and backends, Gateway listeners and routes, pull secrets and role bindings. Deleting one that's still in use says so before you confirm.

Bulk actions

Select rows in any list, even across clusters, and use the bar at the bottom to open one log stream, restart, scale or delete them all. Each write goes to the right cluster, and read-only contexts block it.