Skip to content

Releases

4.0.4

Choose a tag to compare

@guiqui guiqui released this 11 Nov 18:12
5cda859

K8Studio 4.0.4

Status: In preparation
Updated: November 11, 2026

K8Studio 4.0.4 improves Kubernetes 1.37 compatibility, DevSec connection accuracy, topology layout, and YAML and credential-path handling.

Highlights

  • More accurate NetworkPolicy connections across namespaces, including allow-all and deny-all rules.
  • Cleaner DevSec maps: connections follow visible workloads and namespace filters.
  • YAML editing preserves scalar values, including fields named time or creationTimestamp.
  • Device, certificate, and Pod identity views adapt to the APIs available in your cluster.
  • Clearer autoscaling details for zero replicas, scaling tolerances, and scale-to-zero status.

Kubernetes Compatibility

  • Preserved YAML scalar values instead of converting fields based on names such as time or creationTimestamp. Verified kubectl 1.37 KYAML editing and real create/update/delete workflows.
  • Hardened API discovery when preferredVersion is absent and added stable/beta metrics API discovery coverage.
  • Improved HPA displays for minReplicas: 0, scaling tolerances, and ScaledToZero status.
  • Added device resource, certificate, and Pod identity views. Visibility follows the APIs and list operations exposed by the connected cluster, rather than a hard-coded Kubernetes version.
  • Added evidence-based upgrade diagnostics for SELinux, cgroups, node features, and DNS in resource overviews.

DevSec and Visualizer

  • Fixed NetworkPolicy namespace matching across multiple namespaces and prevented unmatched selectors from allowing unrelated workloads.
  • Fixed allow-all interpretation: an empty rule list denies traffic, while an empty rule object allows all peers and ports.
  • Limited DevSec connection lines to rendered canvas workloads, excluding hidden namespaces and owned ReplicaSets that are not separate canvas nodes.
  • Matched selected workloads by ID after object refresh.
  • Improved alignment between connected workloads and services, with coverage for layout stability.

Hardening and Regression Coverage

  • Updated JSONPath and explicitly disabled expression evaluation for credential extraction paths.
  • Removed five unused vulnerable direct dependencies and added build/runtime dependency lockfiles.
  • Added 26 unit regression tests covering policy changes, empty rules, credential paths, and DevSec rendering.
  • Added release CI checks for unit tests, type checking, lint, dependency audits, and packaged startup/native terminal behavior on Windows, Linux, macOS arm64, and macOS Intel.
  • Improved production renderer chunking while retaining application-code obfuscation.

Validation and Availability

The October 16 validation record reports 1,195 passing unit tests, 125 Kubernetes 1.37 checks (including eight live tests), and 23 desktop tests on macOS arm64, covering Kubernetes 1.35 and 1.37. The compatibility checks overlap the unit suite; these counts are not additive. Production builds and the obfuscated macOS test package passed in that validation run.

An Ingress delete/recreate desktop check failed once with CloudMaps active and passed the full rerun; the earlier failure remains recorded for follow-up.

Remaining dependency findings, type/lint failures, cross-platform CI execution, and signed installer, notarization, upgrade, and auto-update validation still block publication. Stable metrics tests use a bridge over real beta metrics samples. Physical device allocation and certificate issuance/rotation are not yet validated.

The website download links remain on the published installer version until 4.0.4 installers are available.

4.0.3

Choose a tag to compare

@guiqui guiqui released this 07 Oct 16:15
5cda859

K8Studio 4.0.3

K8Studio 4.0.3 improves live Visualizer updates, navigation, custom resource editing, and Pod terminal cluster isolation.

Visualizer and Navigation

  • Fixed stale Ingress connections after rules change or are removed (#126).
  • Improved alphabetical workload and Pod ordering, namespace grouping, long resource labels, and persisted namespace filters (#123).
  • Preserved the Visualizer viewport and CloudMaps camera when following resource links and navigating back (#122).
  • Improved Back/Forward navigation and opening resource links and sidebar views in new tabs (#122).

Resource Editing and Terminals

  • Preserved custom resource status when saving resources whose CRD has no separate status endpoint.
  • Fixed opening Prometheus settings from resource Metrics views (#119).
  • Bound Pod exec commands to the selected cluster's kubeconfig, with an explicit context for embedded Pod shells (#111).
  • Cluster-bound terminals now report missing kubeconfig files instead of falling back to the CLI environment (#111).

Regression Coverage

  • Added Playwright checks for Ingress updates, Visualizer filters and layout, and Visualizer/CloudMaps navigation.
  • Verified short and 3,158-line Helm clipboard operations, undo/redo, and saving values to a live test cluster (#113).
  • Added checks for terminating Pod status, namespace filters, closing tabs to the right, and refresh shortcut smoke tests (#112).
  • Added coverage for ConfigMap editing, custom resource status preservation, Prometheus settings navigation, and Pod exec with a conflicting inherited CLI context.

Release Validation

Targeted regression runs passed on macOS arm64 for the Visualizer, CloudMaps navigation, Helm values editing, custom resource editing, Prometheus settings navigation, and cluster-bound Pod terminal flows covered in this release.

Network metrics and node-shell reports from #112 and the original private-CRD crash are not claimed as resolved in 4.0.3.

4.0.2

Choose a tag to compare

@guiqui guiqui released this 17 Aug 07:58
5cda859

K8Studio 4.0.2

Highlights

  • Fixed Security Context reporting so container-level values are resolved first and Pod-level values are used as fallback.
  • Improved MITRE scanner accuracy for non-root execution and privilege escalation findings.
  • Fixed YML tabs that could appear blank for Deployments, Services, RBAC, storage, network, cluster-scoped resources, and CRDs.
  • Made resource updates safer by stripping server-owned status fields before saving.
  • Added local validation for immutable Pod spec updates so unsupported changes are blocked before Kubernetes rejects them.
  • Added Playwright coverage for YAML visibility and updates across representative Kubernetes object kinds and custom resources.
    Fixes
  • Fixed misleading Security Context values that could show the same runAsUser/runAsNonRoot state across different Pods or containers.
  • Fixed Container may run as root findings when effective non-root settings were present.
  • Fixed security-context value colors so safe and unsafe booleans match their severity.
  • Fixed Kubernetes update errors caused by sending object-shaped condition timestamps back to the API server.
  • Fixed detail title status messages overlapping delete, save, and cancel buttons.
  • Fixed YML editor sizing and tab synchronization after opening resources from grids or switching detail tabs.
  • Fixed immutable Pod edit attempts by allowing supported mutable Pod fields and blocking unsupported spec changes with a clearer error.

4.0.1

Choose a tag to compare

@guiqui guiqui released this 06 Aug 09:07
5cda859

v4.0.1

A focused operations and stability release that adds live Caretta traffic to CloudMaps, Deployment traffic analysis, redesigned Helm and RBAC workspaces, expanded privacy controls, and a more reliable K8Studio account portal.

Highlights

  • CloudMaps can now display observed Caretta traffic as map roads between namespaces and workloads.
  • CloudMaps is available during the trial and with Professional or Airtight licenses.
  • Traffic roads use fixed volume bands so traffic levels remain comparable across clusters.
  • Deployment details include incoming and outgoing traffic analysis when Caretta is available.
  • The Helm Package Editor has clearer chart and release workflows, technology icons, and copyable commands.
  • The RBAC Manager organizes subjects and roles into searchable grouped trees.
  • The website and account services include expanded GDPR consent, retention, and account-deletion support.
  • The website now provides direct access to the K8Studio Discord community from its navigation and footer.
  • The account portal has more reliable authentication, responsive layouts, and complete light and dark themes.

CloudMaps Traffic

  • CloudMaps access is limited to active Trial, Professional, and Airtight licenses.
  • Added optional Caretta discovery and automatic local port forwarding for reading observed network flows without exposing Caretta outside the cluster.
  • Added live traffic roads between resolved namespaces and workloads in CloudMaps.
  • Traffic roads show direction and use consistent color and width bands based on absolute traffic volume rather than adapting to the current cluster.
  • Hover details show the source, destination, observed rate, connection count, and destination ports.
  • Traffic routes remain stable while zooming and resolve to the correct map level as the view changes.
  • Viewport-aware traffic rendering and spatial indexing avoid drawing off-screen roads unnecessarily.
  • CloudMaps continues to work normally when Caretta is not installed or temporarily unavailable.

Deployment Traffic Analysis

  • Added a Traffic tab to Deployment details when Caretta is available.
  • Added total, incoming, and outgoing traffic summaries for the selected Deployment.
  • Incoming and outgoing route tables show related workloads or Services, ports, and observed transfer rates.
  • Service-routed traffic is attributed to the relevant Deployment where Kubernetes relationship data is available.
  • Connecting, live, empty, and unavailable states distinguish missing traffic from an application error.

Helm Package Editor

  • Redesigned the Helm workspace with clearer Charts and Releases navigation.
  • Improved chart and release lists with search, counts, versions, descriptions, repository information, and stronger selected states.
  • Added technology icons for recognized charts and releases using the same identification pipeline used by Kubernetes workloads.
  • Improved recognition for common charts such as Argo CD and Metrics Server.
  • Added copy buttons for Helm install commands and code examples in chart documentation.
  • Updated chart details, release details, values editing, and documentation for consistent light and dark themes.

RBAC Manager

  • Redesigned the RBAC workspace to make large sets of subjects, roles, and bindings easier to navigate.
  • Grouped subjects into Users, Groups, and Service Accounts with visible counts.
  • Grouped authorization resources into Roles and Cluster Roles with visible counts.
  • Added searchable, expandable trees instead of presenting every subject or role as one long list.
  • Improved role permissions, assigned roles, bindings, selection headers, and contextual counts.
  • Updated RBAC panels, headers, grids, and editors for consistent light and dark themes.

Website and Community

  • Added a Discord community invitation to the website navigation so help, product discussion, and CloudMaps beta feedback are always within reach.
  • Added the Discord community link and logo to the global website footer alongside the existing social channels.
  • Kept the navigation invitation compact across desktop, tablet, and mobile layouts.

Privacy and Account Portal

  • Expanded website consent controls for necessary storage, analytics, and external media.
  • External videos and embedded media remain blocked until the visitor allows external-media content.
  • Added a persistent privacy-settings entry so consent choices can be reviewed later.
  • Added account-deletion and retention workflows with backend coverage.
  • Restyled licenses, invoices, subscriptions, checkout, activation, invitation, validation, success, and privacy views to match K8Studio.
  • Improved authentication startup, token persistence, session refresh, production caching, responsive layouts, and light and dark themes.

Fixes

  • Fixed Caretta initialization errors when moving between CloudMaps tabs connected to different clusters.
  • Fixed traffic endpoints changing position or disappearing at intermediate zoom levels.
  • Fixed CloudMaps canvas sizing after moving a loading tab into a split view.
  • Fixed Helm chart lists showing the default icon for recognized technologies.
  • Fixed RBAC panels and headers that did not respect light mode.
  • Fixed account portal production login, session refresh, null grid data, invitation error, and checkout-success issues.
  • Fixed account portal sidebar, footer, and grid sizing differences between local and production builds.

4.0.0

Choose a tag to compare

@guiqui guiqui released this 15 Jul 08:05
5cda859

K8Studio 4.0.0

Release date: 2026-07-14

K8Studio 4.0.0 is a major CloudMaps release.

CloudMaps is the Google Maps for Kubernetes clusters: see everything, spot problems, and understand your cluster in seconds. It is designed for large environments, including clusters with 5,000+ workloads, where traditional trees and tables are not enough on their own.

This release also adds first-class Gateway API visibility, redesigned workload and Pod details, clearer workload status classification, improved port-forward management, and timeline polish.

Highlights

  • CloudMaps is now the center of the K8Studio 4.0 experience, giving teams a visual way to explore Kubernetes relationships, routing, dependencies, and infrastructure context.
  • Gateway API resources now have their own navigation area, grids, and detail views for modern Kubernetes ingress and routing workflows.
  • Gateways, GatewayClasses, HTTPRoutes, GRPCRoutes, and ReferenceGrants are easier to inspect with focused overview pages instead of raw YAML-first views.
  • Workload status is more precise with Healthy, Scaling, Updating, Pending, Blocked, and Degraded states, including better handling for Pending Pods.
  • Deployment and Pod details use space more efficiently with redesigned status, strategy, resource, container, labels, annotations, ports, and policy sections.
  • Port forwarding is easier to manage from resource actions with active-state indicators, inline creation, and quick removal.
  • Timeline event labels behave better in dense rows, avoiding occasional overlap between lifecycle labels.

CloudMaps

  • Added CloudMaps as the headline K8Studio 4.0 experience: a Google Maps-style way to monitor and explore Kubernetes clusters.
  • Improved visual investigation for huge clusters, including environments with 5,000+ workloads.
  • Made Kubernetes relationships easier to follow across selected objects, workloads, Services, Pods, Nodes, ports, routes, policies, storage, and dependencies.
  • Integrated Gateway API resources into the CloudMaps troubleshooting flow so modern ingress and routing resources can be understood in context.
  • Moved Topology after Logs and before Events in object details for a more natural investigation flow.
  • Improved dense timeline readability so CloudMaps investigations are easier to correlate with recent cluster activity.

Gateway API

  • Added a dedicated Gateway section in the cluster tree when Gateway API resources are available in the cluster.
  • Added grids and resource views for Gateways, GatewayClasses, HTTPRoutes, GRPCRoutes, ReferenceGrants, BackendTLSPolicies, and BackendTrafficPolicies.
  • Gateway details now highlight gatewayClassName, listeners, listener ports, addresses, and conditions.
  • GatewayClass details now show controller information, parameters, and conditions without forcing users into raw YAML first.
  • HTTPRoute and GRPCRoute details now show hostnames, parent refs, route rules, matches, filters, backend refs, ports, weights, and route conditions.
  • ReferenceGrant details now show allowed From and To references in a clean overview.
  • Gateway condition types are color-coded consistently with the rest of the workload and resource status experience.
  • Gateway API resources use a dedicated Gateway icon so they are easier to distinguish from general networking objects.

Workload Status

  • Added clearer customer-facing workload states: Healthy, Scaling, Updating, Pending, Blocked, and Degraded.
  • Deployments with non-crashing Pending Pods are shown as Pending instead of being marked Degraded too early.
  • Scheduling failures, progress-deadline failures, and blocked rollouts are shown as Blocked.
  • Crash loops, ReplicaFailure, and hard container errors continue to appear as Degraded.
  • Status colors now match the existing Kubernetes object icon and status palette.
  • Replica chips now include a Pending state so status cards explain rollout state more clearly.

Resource Details

  • Deployment details now make better use of horizontal space with compact Status, Strategy, Resources and Limits, and Pods sections.
  • Deployment Pod grids use available width and height more efficiently while avoiding large empty areas.
  • Pod details now start with a compact Pod Details card that groups status, conditions, scheduling, timing, and access information.
  • Labels and annotations are moved lower in the page so operational status, resources, ports, policies, and containers are easier to reach first.
  • Container cards now sit side by side with Network Policies and Port Forwarding when space allows.
  • Container information popovers are more readable, with clearer labels, values, commands, ports, environment variables, probes, security context, and mounts.
  • Conditions are now labeled as Conditions instead of Health, and condition types are color-coded.
  • Light-mode detail buttons have stronger contrast and no longer look disabled.

Port Forwarding

  • Port forwarding now uses a clearer two-arrow action icon with an active-state dot.
  • The port-forward action shows whether a resource has active port forwards with a subtle green or grey indicator.
  • Hovering the action opens a compact popover with active sessions, local URLs, remote/local port information, and remove controls.
  • New port forwards can be created inline from the popover when none are active.
  • The detail-page Port Forward box now renders only when port forwarding exists and stays compact beside related networking information.
  • Long active port-forward lists are scrollable so the surrounding detail layout stays stable.

Timeline and Topology

  • Timeline duration labels now avoid overlapping in dense rows such as Started, Created, Pulled, and SuccessfulCreate event clusters.
  • Extra colliding timeline markers are summarized behind compact counts instead of stacking labels on top of each other.
  • Narrow duration markers now render more compactly so event labels do not collide with count badges.
  • Topology tabs now sit after Logs and before Events in object details.

Fixes

  • Fixed Service details that could crash when rendering object-shaped status fields such as loadBalancer.
  • Fixed missing component views for newly supported Gateway API resources.
  • Fixed Gateway, GatewayClass, HTTPRoute, GRPCRoute, and ReferenceGrant detail pages falling back to raw or missing views.
  • Fixed workload status wording and color mismatches when a Deployment had Pending Pods.
  • Fixed occasional timeline label overlap in dense event rows.
  • Fixed several layout issues in Pod and Deployment details where cards, grids, or long container paths could waste space or overflow.

3.1.10

Choose a tag to compare

@guiqui guiqui released this 03 Jul 10:02
5cda859

K8Studio 3.1.10

Release date: 2026-07-03

A focused Visualizer release with smoother zooming and panning, clearer namespace containers, better cross-namespace links, and sharper high-zoom rendering.

Highlights

  • Zooming and panning are smoother and avoid whole-window flicker during repeated Visualizer interactions.
  • High-zoom element rendering is sharper without using an excessive cache scale.
  • Maximum zoom is aligned with the cached image scale so zoom-out does not get stuck beyond the useful viewport limit.
  • Namespace and subgroup containers are visible again with simpler styling and clearer dark-mode contrast.
  • Cross-namespace Service and Ingress links are handled more consistently.
  • Filtered views now hide links to namespaces that are not currently visible.

Visualizer

  • Zooming and panning are smoother and avoid the whole window flicker that could appear during repeated interactions.
  • Visualizer caching now keeps high-zoom element images sharper without using an excessive render scale.
  • Maximum zoom is aligned with the cached image scale so zoom-out does not get stuck beyond the useful viewport limit.
  • Namespace and subgroup containers are visible again with simpler styling and clearer dark-mode contrast.
  • Cross-namespace Service and Ingress links are handled more consistently.
  • Filtered views now hide links to namespaces that are not currently visible.
  • Added focused Visualizer and positioner tests around cached resets, getter-only dimensions, malformed links, filters, zoom cache scale, and group versions.

Rendering and Layout

  • Visualizer elements remain more readable at the highest zoom level.
  • Cached element images use a balanced scale for crisper text without the memory cost of an oversized cache.
  • The zoom limit now matches the rendered cache scale so users can zoom back out reliably.
  • Namespace containers and subgroup containers use simpler styling without heavy shadows or gradients.
  • Subgroup backgrounds are slightly darker so nested areas remain distinguishable.

Fixes

  • Fixed Visualizer links remaining visible when their namespaces were filtered out.
  • Fixed namespace and subgroup boxes disappearing after simplifying the Visualizer styling.
  • Fixed a cached layout reset issue that could try to write dimensions to getter-only position objects.
  • Improved handling for malformed Visualizer links so layout calculation remains stable.
  • Added focused tests for Visualizer cache scale, namespace filtering, group versioning, graph reset behavior, and positioner cache behavior.

3.1.9

Choose a tag to compare

@guiqui guiqui released this 30 Jun 12:35
5cda859

K8Studio 3.1.9

Version 3.1.9
A release focused on a more flexible AI Copilot, visual Kubernetes Topology, and a new Events Timeline for understanding what is happening inside a cluster faster.

Release date: June 28, 2026

Highlights

  • Copilot can now work with more AI providers, including local CLI-based assistants and API-key based providers.
  • Copilot provider and model selection is easier to manage from settings and can be changed without closing the Copilot window.
  • New Topology views help you see how a selected Kubernetes object connects to related workloads, Services, ConfigMaps, Secrets, storage, RBAC, and Nodes.
  • New Events Timeline turns Kubernetes Events into a visual, searchable history grouped by resource hierarchy.
  • Warnings and failures are easier to spot with clearer event markers, density signals, and compact counts.
  • Timeline and Topology are available for Professional and trial users.

AI Copilot

  • Choose from available AI providers in one compact settings list.
  • Use installed CLI-based assistants such as Claude Code, Codex, Gemini, GitHub Copilot, and Ollama when they are available on your machine.
  • Configure OpenAI and Claude with API keys when you prefer a hosted provider.
  • Switch provider and model directly from Copilot without closing the current conversation.
  • Provider icons, default selection, unavailable states, and setup guidance are clearer and easier to scan.
  • When a provider needs to be installed or logged in, K8Studio now gives clearer setup feedback.

Topology

  • Resource details now include a Topology tab for visually exploring related Kubernetes objects.
  • See common relationships across workloads, Pods, Services, Ingresses, ConfigMaps, Secrets, ServiceAccounts, PVCs, PVs, StorageClasses, HPAs, PDBs, NetworkPolicies, RBAC bindings, and Nodes.
  • Pan, zoom, and center the diagram while keeping the selected object easy to find.
  • Kubernetes object icons and animated connected arrows make relationships easier to follow.
  • Shared infrastructure objects such as kube-root-ca ConfigMaps or busy Nodes stay focused unless you select them directly.
  • Topology keeps noisy clusters readable by focusing the diagram on the objects most relevant to the selected resource.

Events Timeline

  • Added a new Kubernetes Events Timeline for exploring cluster events visually.
  • Events are grouped by resource hierarchy so App, Deployment, ReplicaSet, Pod, Node, Job, and CronJob activity is easier to scan.
  • Instant events render as compact diamonds while duration events render as real time spans, making warnings, BackOff loops, FailedMounts, scheduling, image pulls, and lifecycle activity easier to distinguish.
  • Added density strip, now-line, readable time axis, compact event counts, namespace/kind/severity filters, search, time-range chips, tooltips, collapsible rows, per-row sparklines, and a resizable resource column.
  • Timeline supports dark and light themes, horizontal panning, warning glow states, readable resource names, and namespace labels under resource names.
  • Busy clusters remain easier to navigate because dense event streams are summarized instead of overwhelming the screen.

Polish and Plans

  • Timeline and Topology are available on Professional plans and during trials.
  • Warning and failure events are now easier to identify at a glance.
  • Timeline controls use less vertical space, leaving more room for the event lanes.
  • Long resource names and namespaces are easier to read in the timeline resource column.
  • Timeline panning and resource-column resizing feel smoother during day-to-day investigation.

3.1.8

Choose a tag to compare

@guiqui guiqui released this 26 Jun 06:44
5cda859

K8Studio 3.1.8 - 2026-06-25

Highlights

  • Added cluster visibility controls so users can hide clusters from large kubeconfig files and bring them back from a dedicated hidden-clusters modal.
  • Improved cloud authentication testing and documentation for EKS exec, AWS SSO-backed EKS, GKE with gke-gcloud-auth-plugin, AKS with kubelogin, and Keycloak OIDC.
  • Added a local observability lab path for kind clusters with Prometheus, Grafana, and Argo CD.
  • Improved cluster overview, workload overview, and grid loading states with animated skeletons.
  • Improved Visualizer behavior so automatic layout runs when resources are added or removed, while preserving manual layouts for normal updates.
  • Added a collapsible Visualizer property editor with better button placement and persistence.
  • Added integrated and external terminal preferences, including a default terminal selection path.
  • Improved provider detection icons so EKS, GKE, and AKS clusters use AWS, Google Cloud, and Azure branding automatically.
  • Updated the website homepage, feature carousel, feature screenshots, download CTA, SSO resources, AI assistant page, and showcase image performance for the 3.1.8 release.

Cluster Management

  • Added support for hiding clusters imported from the default kubeconfig or any additional kubeconfig file.
  • Added separate available and visible cluster concepts so users can import many clusters while keeping the tree focused.
  • Added a cluster tree menu action to hide a cluster.
  • Changed the cluster add button hidden-cluster picker into a modal so large kubeconfig files remain manageable.
  • Preserved hidden cluster state until the user explicitly makes a cluster visible again.
  • Added test kubeconfig coverage for fake EKS and GKE contexts so cloud provider icon behavior can be validated locally.

Authentication and Cloud Providers

  • Added and documented a Keycloak OIDC local test flow for kubeconfig OIDC login.
  • Added AWS SSO-backed EKS kubeconfig support using the AWS exec credential plugin.
  • Added GKE kubeconfig support using the gke-gcloud-auth-plugin exec credential plugin.
  • Added AKS kubeconfig support using kubelogin.
  • Improved guidance for cloud SSO login behavior so K8Studio can surface missing external credential tools more clearly.
  • Improved cluster icons for managed clouds: EKS now uses AWS styling, GKE uses Google Cloud styling, and AKS uses Azure styling.

Visualizer

  • Changed Visualizer auto-layout behavior so it relayouts only when graph elements are added or removed.
  • Preserved manual node positioning during normal refreshes and data updates.
  • Added a collapsible property editor for selected Visualizer items.
  • Improved collapse and expand button placement so it no longer overlaps the property editor title, icons, or DevSec toggle.
  • Kept the property editor collapsed across subsequent selections until the user expands it again.
  • Aligned Visualizer buttons with the existing scan and toolbar button styling.

Loading States and Resource Views

  • Added animated skeletons for cluster overview while data is loading.
  • Added animated skeletons for workload overview while data is loading.
  • Added loading polish for grid-based resource views.
  • Improved empty-looking loading states so slow cluster calls are easier to distinguish from empty data.

Terminal

  • Added settings for integrated versus external terminal behavior.
  • Added a default terminal selection path.
  • Improved shell fallback handling for containers that do not provide bash, sh, or a shell in PATH.

3.1.7

Choose a tag to compare

@guiqui guiqui released this 22 Jun 11:35
5cda859

Version 3.1.7

A Copilot and reliability release with safer analyzer-backed fixes, stronger privacy controls, better large-cluster metrics handling, improved license onboarding, and more polished web integrations.

Release date: June 19, 2026

Highlights

  • Copilot can now use K8Studio tools to open views, inspect resources, analyze issues, and prepare reviewed Kubernetes changes.
  • Fix flows are safer: Copilot diagnoses first, uses analyzer-built fixes, shows a before/after YAML diff, runs preflight checks, and waits for approval.
  • Analyzer coverage is broader with real-world troubleshooting scenarios, golden tests, and richer evidence from related resources, events, logs, status, and metrics.
  • Sensitive data sent to LLM providers is anonymized more aggressively, including names, namespaces, labels, annotations, images, env vars, secrets, URLs, IPs, domains, and logs.
  • Large-cluster metrics handling is more efficient, with deduped metric fetching and unchanged-update skipping for workload CPU and memory columns.
  • Metrics views are easier to explore with full-width charts, per-chart refresh, bottom legends, and cleaner loading states.
  • macOS shutdown is more reliable after fixing a native filesystem watcher cleanup crash that could appear when quitting the app.

Copilot and Analyzer

  • Added K8Studio tool support so Copilot can open clusters, pages, resource grids, selected resources, logs, YAML, events, Helm, Security, Metrics, Visualizer, Argo CD, Prometheus, and Grafana views.
  • Copilot now understands more natural requests such as opening Pods, Deployments, Namespaces, Events, Leases, Services, Ingresses, ConfigMaps, Secrets, RBAC resources, and other Kubernetes views, including common misspellings and singular/plural forms.
  • For fix requests, Copilot now runs the analyzer first and uses only analyzer-built executable fix candidates for automatic changes.
  • Fix proposals now open a reviewed before/after YAML diff with synchronized scrolling, editable proposed YAML, a clear explanation, confidence and risk labels, and a no-safe-automatic-fix state when needed.
  • Mutation flows now run RBAC and server-side dry-run preflight before applying changes and always wait for user approval.
  • Added stricter Copilot behavior rules to prevent invented Kubernetes fields, impossible live Pod patches, unsafe kubectl instructions, and model-authored mutation payloads when an analyzer fix is available.
  • Added a concise Copilot welcome message and a shared help modal that explains what Copilot can do now.

Analyzer Coverage and Confidence

  • Expanded per-kind analyzer reasoning for Pods, Deployments, ReplicaSets, StatefulSets, DaemonSets, Jobs, CronJobs, Services, Ingresses, Endpoints, PVCs, PVs, StorageClasses, HPAs, PDBs, Quotas, LimitRanges, RBAC, and ServiceAccounts.
  • Improved evidence collection so analysis can use the current object, parents, owned children, related events, recent logs, status, conditions, metrics, and known failure signatures.
  • Added analyzer support for safer fix candidates around common workload, service, storage, quota, limit range, scheduling, image, command, and readiness issues.
  • Added a real-world troubleshooting scenario fixture set with 35 known Kubernetes failure cases.
  • Added analyzer golden tests, scenario coverage tests, MCP contract tests, LLM response parser tests, dry-run/preflight tests, and regressions for bad LLM output.
  • Improved handling when no safe automatic fix can be built from analyzer evidence, so Copilot explains what is missing instead of inventing a patch.

Privacy and Safety

  • Improved LLM privacy with reversible local anonymization for names, namespaces, labels, annotations, image references, environment variables, secret references, URLs, IP addresses, domains, and log text.
  • Secrets and secret-like values are redacted before context is sent to an LLM provider.
  • Added additional protection for free-form evidence and logs, including JWT-shaped tokens, cloud access keys, prose token values, IPv4 and IPv6 addresses, emails, URLs, domains, and image references.
  • Added tests proving sensitive secret payloads and common credential formats are not leaked through Copilot context.

Metrics and Large-Cluster Performance

  • Metrics pages now use wider chart cards, bottom legends, and an expanded modal view for inspecting individual charts.
  • Added per-chart refresh behavior so refreshing one metric chart no longer reloads the full metrics page.
  • Improved chart loading states with a clearer spinner and loading text.
  • Improved CPU chart readability by separating CPU capacity styling from CPU utilization.
  • Improved workload CPU and memory columns for Deployments, StatefulSets, DaemonSets, and ReplicaSets.
  • CPU and memory metric cells now use clearer rounded values with smaller unit text.
  • Resource usage graphs now show only when request or limit data is available.
  • Metric refresh is more efficient through deduped requests and skipped UI updates when metric values have not changed.
  • Metric refresh behavior can be controlled from cluster settings.
  • Improved cluster overview resource usage cards for CPU, memory, requests, limits, allocatable capacity, and refresh behavior.

Resource Details and Status

  • Pod status is now calculated from the same shared logic across grids, details, and Deployment overview Pod components.
  • Deployment overview Pod status now better matches the attached Pods instead of showing misleading warnings.
  • Pod component links in Deployment overview now open the related Node correctly.
  • Container volume mount cards expand better for long paths and avoid text overflow.
  • YAML views now include a copy button for the current manifest content.
  • Default cluster icons now use the K8Studio SVG mark and tint automatically with the cluster's configured color.

License, Trial, and Onboarding

  • Improved license and trial display so days remaining come from the backend days-left value.
  • Fixed stale license UI state when an expired license is refreshed and replaced by a valid license.
  • Removed misleading trial-start messaging from the license activation window.
  • Reworked license activation flows so buying Pro, entering a license code, online activation, and Airtight activation are clearer and separated.
  • Added trial-days remaining display in the bottom bar with clearer color states.
  • Added EULA support and improved onboarding/legal consent presentation.

Website and Download Tracking

  • Added backend and website support for download tracking with platform and device information.
  • Added cookie consent messaging for website visitors.
  • Fixed CORS configuration for download telemetry requests from k8studio.io.
  • Improved first-five-minute onboarding messaging and license prompts for new users.

Web Integrations

  • Argo CD, Grafana, and other web integration pages keep passwords hidden by default with a reveal control.
  • Credential bars now provide separate copy buttons for username and password.
  • Embedded web integration pages continue to use port forwarding to open local management UIs when available.

Fixes

  • Fixed a macOS quit crash caused by native filesystem watchers being torn down while Electron was shutting down.
  • Fixed a crash when clicking links to resources that no longer exist.
  • Fixed several Copilot tool-routing issues where requests such as opening Deployments, Helm, Security, Nodes, Namespaces, Events, Leases, and other pages were not recognized reliably.
  • Fixed Copilot handling for bad or incomplete model-generated patch payloads.
  • Fixed a readiness-fix flow where Copilot could not build a safe patch from analyzer evidence.
  • Fixed a UI issue where the Copilot new-chat button looked like a no-op on empty conversations.
  • Fixed Copilot welcome content that was too large by replacing it with a compact summary and shared help modal.
  • Fixed website desktop breakpoints so large desktop screens do not incorrectly show the mobile navigation.
  • Fixed the Next.js SWC binary issue for local website development on darwin/arm64.

3.1.6

Choose a tag to compare

@guiqui guiqui released this 09 Jun 19:43
5cda859
v3.1.6

Adding ssh and multicluster