Your Kubernetes fleet,from the first VM to the last byte.
Talos Linux clusters on Proxmox VE — created, scaled, upgraded, repaired and deleted for you. Safely, observably and live, across every environment.
Everything in one place. One click away.
Each tile opens the real page — with your live data.
Fleet overview
Every environment at a glance: health, clusters, live operations and trends.
Clusters
Cards or table, states, versions, favourites and bulk actions.
Operations
Step-by-step timelines, retries, resume and abort — streamed live.
Create a cluster
A guided wizard with capacity checks, templates and an optional lifetime.
Hosts & disks
Cordon, drain and SMART health of every disk before it fails.
Metrics
30 days of history: latency, errors, reconcile lag and host memory.
Alerts
Rules with thresholds, push and Discord — with quiet hours.
Network & addresses
Every allocation, IP search and how many clusters still fit.
Storage
Ceph health, pools, per-cluster RBD/CephFS/S3 and when it fills up.
Version matrix
Talos, Kubernetes and every add-on against the catalog — and the upgrade path.
Schedules
Nightly drift checks, extra etcd snapshots and upgrades in a maintenance window.
System & backups
Backup checklist, API compatibility, TLS and notifications.
Audit log
Who did what, when — down to the signed-in person.
API tokens
Scopes, expiry reminders and a guided rotation of the console's own token.
Environments
Lab, staging and production side by side, production marked red.
One request. Five careful phases.
Every operation runs preflight → perform → verify → commit, survives restarts and stops safely when something is off.
- 1
Plan
Spec, capacity, address space and versions are checked before anything is touched.
- 2
Provision
Proxmox VMs, a private SDN network, DHCP and the gateway pair — owned and labelled.
- 3
Bootstrap
Talos boots, etcd forms exactly once, Kubernetes answers through the gateway.
- 4
Equip
Cilium, MetalLB, cert-manager, Ceph CSI and ingress, pinned to tested versions.
- 5
Operate
Scale, upgrade by replacing VMs, repair drift, back up — and delete with a plan.
Not refreshed. Live.
TFM streams every change the moment it happens. The console updates in place — and your phone buzzes when an operation finishes, fails or stops.
- Server-Sent Events from TFM, shared by every tab
- Installable app with push notifications, even when closed
- Quiet hours and failures-only, per device
Careful by design.
Delete with a plan
You see everything a delete removes, then confirm with a short-lived token.
Idempotent & resumable
Every step is recorded; a crash resumes from the same step, never twice.
Roles & single sign-on
Viewer, operator, admin — TFM's audit names the person, not just a token.
Secrets stay sealed
Tokens and keys are encrypted at rest and never reach the browser.
Honest about HA
A single host is shown as a single point of failure — everywhere.
Backups you can trust
A checklist shows what is protected and what is not.