Multi-cluster, one view
Enable any set of kubeconfig contexts: pods, workloads, nodes, and events merge into a single table, every row tagged with its cluster. One unreachable cluster never blanks the rest.
Native · iPhone, iPad & Mac · On the App Store
One live view of every cluster you care about — pods, workloads, logs, GitOps and capacity — native on the Mac and a full dashboard in your pocket. Widgets and Siri tell you when something breaks, and a week of history shows what you reserve against what you use. It never writes to your cluster.
And at your desk — the Mac app
Real screenshots throughout — captured from the apps running in demo mode. Native Swift on iPhone, iPad, and Mac.
Why native matters. Electron dashboards ship a browser to draw a table. KubeGlance is Swift against Apple's own frameworks: it launches instantly, idles quietly on battery, and renders 1,000-pod tables at full frame rate.
Kubeconfig in. Nothing out. Your kubeconfig's certificates and tokens, your cluster's CA checked on-device. No in-cluster agent, no account, no telemetry — and read-only by default, so it can stay open against production.
Is my cluster okay? What are my pods doing? What changed? This is the loop you run every day — the platform views that answer the harder questions come next.
Enable any set of kubeconfig contexts: pods, workloads, nodes, and events merge into a single table, every row tagged with its cluster. One unreachable cluster never blanks the rest.
Type namespace:prod status:crash and completed filters become chips. The field suggests live values — your real namespaces, statuses, and nodes — as you type.
Deployments, StatefulSets, DaemonSets, and Jobs with the one number that matters — ready over desired. Healthy rows stay quiet; only trouble gets color.
Every Deployment keeps its rollout history. Expand a past revision to see exactly what that deploy changed — image bumps, resource edits, containers added or removed.
Live CPU and memory from metrics-server. Prometheus, VictoriaMetrics or Thanos is found on every cluster by itself, and history arrives through the API server's service proxy — no port-forwarding, no exposure, no settings.
Merge any pods — across workloads, namespaces, even clusters — into one chronological stream. Structured JSON and nginx lines parse into fields; bookmark lines and jump between them.
Crash loops, NotReady nodes, stalled rollouts, expiring certificates — and the ones you want before an upgrade window, like a drain a budget would hang. Native notifications, scoped by cluster and namespace; critical ones break through Focus. How alerts work on each device.
Connect Amazon EKS and Google GKE with the credentials you already have — no aws, gcloud, or exec plugins, which can't run on iPhone anyway. Pick the cluster by name; KubeGlance signs in natively. Azure AKS and OIDC too.
Cluster health, recent releases and load on the Home Screen, Lock Screen and Mac desktop — per cluster or per namespace. Ask Siri to check a cluster, run it from Shortcuts, and tap an alert to land on the pod it is about.
Lists tell you what exists. These screens tell you what it means — before the upgrade window, when a promote stalls, and when the bill arrives.
Capacity
KubeGlance compares what each node has promised — requests against allocatable, the scheduler's own arithmetic — and says in one sentence whether the rest of the cluster absorbs the loss. Then it shows where every replica actually sits, from the pods that exist rather than the spec.
GitOps · Flux
Flux says DependencyNotReady but not which dependency. KubeGlance draws the chain — sources, then every Kustomization and Helm release in dependency order — and names the object that broke and how many wait behind it. Waiting is amber, broken is red, so you fix the cause, not the symptom.
Efficiency
The scheduler counts requests, not usage, so a cluster can be full on paper and idle in fact. Efficiency splits each cluster's CPU and memory into in use, reserved but idle, and never reserved — averaged over 24 hours or 7 days from the Prometheus, VictoriaMetrics or Thanos it finds on its own.
Widgets · Siri
Health, releases and load on the Home Screen, Lock Screen and Mac desktop, per cluster or per namespace. The health widget changes colour only for what is broken now — a budget that has blocked drains for a month is a quiet "3 advisories", not a warning you learn to ignore.
Platform views
Each appears only when your cluster has the capability — Flux, Argo CD, KEDA, Flagger, cert-manager, Gateway API are detected through the API server, never assumed. A plain cluster stays a plain cluster.
Sync and health as two answers — drifted but fine, deployed and broken. Drift says whether self-heal reverts it; app-of-apps link parent to children.
Kustomizations, Helm releases and sources with applied versus attempted revision, and the dependency that blocks a reconcile.
Every workload's image tag, when it rolled out, and the Flux object that manages it — git SHAs become commit links.
HPA and KEDA merged per workload, attention states in words, replicas and each metric charted against its target.
Flagger rollouts: the phase, how much traffic moved, and failed checks against the rollback threshold.
Per container, usage against the request and the limit on one bar — the one about to be OOM-killed is obvious.
Claims that never bound, volumes bound to nobody, and namespaces at their quota — the Pending a node view cannot explain.
Which device, on which node, held by which pod — and whether a waiting claim is a shortage or a mismatch.
Certificates cert-manager and TLS secrets sorted by days left, stuck issuance visible
Network Services with their real endpoints, Ingress and Gateway API routes with per-parent health
Config, CronJobs, node pools Pods running stale config, last-run verdicts, capacity against the provider's ceiling
iPhone & iPad
Kubernetes GUIs stop at the desktop. The iPhone and iPad app is the same dashboard on the same Swift core — every view above, not summaries of them. When the pager goes off at dinner, triage from your pocket.
Lens made Kubernetes visual. KubeGlance keeps the loop you love and removes the weight around it.
| KubeGlance | Lens | |
|---|---|---|
| Platforms | macOS, iPhone and iPad — native Swift | macOS, Windows, Linux |
| Account required | Never | Lens ID activation |
| Changes to the cluster | Read-only by default | Resource editor, terminal, port forwarding |
| Multi-cluster | One merged view, every row tagged with its cluster | Multi-cluster management |
| GitOps | Flux dependency chain and Argo CD sync vs health, built in | Helm; extensions |
| Capacity & efficiency | Node-loss verdict, drain preview, reserved vs used, suggested requests | Metrics and logs |
| iPhone / iPad | Full dashboard app, same core | No iOS app listed |
| Source code | Closed | Closed; the open-source Lens was retired |
| Price | Free on Mac · Pro on iPhone & iPad (3-day free trial) | Free Personal tier; paid plans required above $10M revenue or funding |
Checked against Lens's own pricing and docs on 25 September 2026. Detailed comparisons: Lens · k9s · Headlamp · Aptakube · Kubernetes apps for iPhone.
Yes — this is KubeGlance's unusual trick. The iPhone and iPad app is a real dashboard on the same core as the Mac app: overview, pods, workloads, events, and merged logs, with identical multi-cluster merging and read-only guarantees. Most Kubernetes GUIs are desktop-only.
Yes. Argo CD is detected through the API server, and Applications are listed with sync and health as two separate columns — Argo answers two different questions with them. An app can be Synced and Degraded (deployed faithfully, and broken) or Out of sync and Healthy (changed by hand, and fine for now), and those need opposite responses. Drift says whether self-heal will revert it or it waits for a manual sync; app-of-apps parents link to their children; apps an ApplicationSet generated are marked as such. Flux works the same way, so a cluster mid-migration shows both.
Yes — it covers the everyday Lens workflow (kubeconfig in, overview, pods, workloads, logs, events) as a native app for Mac, iPhone and iPad. It is read-only by default, free on the Mac, and needs no account. How the two compare.
No. KubeGlance talks to the Kubernetes API with your existing kubeconfig credentials. No agent, no operator, no telemetry.
Client certificates (mTLS) and bearer tokens, including inline PEM and multi-file kubeconfigs with kubectl's merge semantics — plus native login to Amazon EKS, Google GKE, Azure AKS, and generic OIDC providers. No exec plugins, so cloud clusters work on iPhone too.
Yes. KubeGlance implements the AWS and Google sign-in handshakes itself — it signs the STS request and mints the Google token on your device — so there's no exec plugin to install and it works on iPhone and iPad. Enter your credentials (AWS keys or SSO, a Google service-account key or Sign in with Google) and pick the cluster from the list. Nothing is sent to us; credentials stay in the device keychain.
Yes. KubeGlance detects them through API discovery and adds the GitOps, Autoscaling, and Certificates views — with health derived from each object's own conditions the way Flux and Argo do it. Sections appear only on clusters that have the capability, so a plain cluster stays uncluttered. Argo CD is covered too.
Yes. Alert rules cover crash loops, restart storms, NotReady nodes, stalled rollouts, failed jobs, expiring certificates, services without endpoints, and pods near their memory limit — scoped by cluster color, namespace, and name patterns. The Mac app stays resident and notifies within seconds; on iPhone and iPad, checks run whenever iOS grants background time. Details per device.
No. Live numbers come from metrics-server. If Prometheus, VictoriaMetrics or Thanos runs in the cluster, KubeGlance finds it by itself and adds history charts and 24-hour or 7-day efficiency figures, querying it through the API server's service proxy — even minimal setups work via kubelet resource metrics or app-level process metrics.
Not today, by design. KubeGlance is read-only by default — it watches, streams, and inspects.
It's on the App Store — one listing for iPhone, iPad, and Mac. The Mac app is free; iPhone and iPad come with a 3-day free trial of KubeGlance Pro.