Skip to main content
When you sign in you want to know what broke overnight, what is waiting on you and which cluster to open first. The Dashboard answers that for the organisation you have selected, from top to bottom in that order.
The Dashboard only shows clusters in the organisation you currently have selected. Switch organisations from the profile menu or the command palette (Switch Organisation) to see a different fleet.

Briefing

The line at the top summarises what changed since your last visit - for example operations that ran or failed, plans approved, AI sessions that completed and how cloud spend is pacing this month. Each sentence links to where it came from. When nothing needs you it says so. A new organisation also sees a setup checklist here.

Your clusters

Quick links to the clusters you subscribe to, so the ones you work on every day are one click away. Browse clusters opens the full cluster list.

Needs your attention

Approvals, failures and alerts waiting on you, in one place:
  • Inbox - the same personal worklist as Activity & Inbox: plans to approve, actions to confirm, failed steps and AI sessions waiting for an answer.
  • Active alerts - alerts firing across the fleet. See Alerts.
  • Fleet issues - offline clusters, failed reconciles and unhealthy add-ons, most severe first. Diagnose with AI hands an issue to the AI Assistant with its context attached.

Fleet overview

Everything running across the organisation:
  • Fleet globe - each cluster at its real location. Tower height follows node count and colour follows live status; select a cluster to open it.
  • Fleet pulse - clusters online, GitOps health, failing resources, and cloud spend with the projected month-end total by provider and by top cluster. The full breakdown is in Cloud Cost.
  • Fleet Security - actionable critical and high findings across the fleet, with the worst cluster. Opens the Security Center.
  • Failure domains - which clusters survive losing a zone and which lose quorum.
  • Fleet Backups - backup posture and the worst cluster, when backups are enabled for your organisation.
  • CPU and memory across the fleet.

Clusters missing from the globe

Ankra places a cluster from its nodes’ topology.kubernetes.io/region label first, and otherwise from the public IP the agent reports, which gives an approximate location. A cluster with neither is listed separately instead of on the globe: Draft and archived clusters never appear on the globe.

Next step

Open the first item under Needs your attention. If it is a failed deploy, find out why it failed and retry it.