sarman/tftsr-devops_investigation
1
1
Fork 0
Code Issues 4 Pull Requests Actions 34 Packages Projects 3 Releases 6 Wiki Activity
aadde0f336
tftsr-devops_investigation/TICKET-kube-action-namespace-stability.md
Shaun Arman 5f4ca1291a
All checks were successful
Test / frontend-tests (pull_request) Successful in 1m40s
Details
Test / frontend-typecheck (pull_request) Successful in 1m49s
Details
PR Review Automation / review (pull_request) Successful in 4m12s
Details
Test / rust-fmt-check (pull_request) Successful in 11m49s
Details
Test / rust-clippy (pull_request) Successful in 13m6s
Details
Test / rust-tests (pull_request) Successful in 14m36s
Details
docs: add ticket summary for kube action namespace and stability fixes
2026-06-08 22:04:53 -05:00

5.6 KiB
Raw Blame History

Ticket Summary — Kubernetes Action Namespace & Stability Fixes

Branch: fix/kube-action-namespace-and-stability PR: #86

Description

Seven bugs in the Kubernetes management interface were identified via systematic debugging and resolved across 6 commits.

The most severe was a temp kubeconfig race condition in the Rust backend: every kubectl-based IPC command wrote a temp file to a static path derived only from cluster_id. Concurrent calls — triggered by rapid section or namespace switching — shared identical paths. TempFileCleanup::drop() on the first-to-finish call deleted the file while a concurrent kubectl process was still reading it. Errors were silently swallowed, leaving the UI showing stale/empty data. This was the root cause of "things stop loading after a few selection changes."

The second major class of bugs was namespace "all" passed to targeted kubectl commands. When the user selects "All Namespaces", KubernetesPage stores selectedNamespace = "all" and passes it as a prop to all list components. loadResourceData correctly converts "all" → "" for list fetching (which becomes --all-namespaces in Rust). However, action handlers inside list components (edit, delete, scale, logs, shell, attach) used the raw prop and forwarded "all" to kubectl -n all, producing "namespaces 'all' not found" errors.

Acceptance Criteria

  • Rapid section/namespace switching no longer causes data to stop loading
  • Pod Logs loads successfully when "All Namespaces" is selected
  • Pod Shell, Attach, and Edit open and target the pod's actual namespace
  • Deployment, StatefulSet, DaemonSet, and all other workload action commands work under "All Namespaces"
  • Network, Config, Storage, and Access Control action commands work under "All Namespaces"
  • Workloads → Overview shows actual resource counts (not all-zero)
  • Cluster connection errors display a visible banner instead of failing silently
  • connectClusterFromKubeconfigCmd is only called once on mount, not twice
  • Dark mode — all text is readable; status indicators are visible

Work Implemented

Commit 1 — fix(kube): unique temp kubeconfig paths

File: src-tauri/src/commands/kube.rs

Added KUBECONFIG_COUNTER: AtomicU64 and unique_kubeconfig_path(cluster_id) helper. Replaced all 74 static temp_dir.join(format!("kubeconfig-{}-*.yaml")) calls with the helper. Each invocation now gets a globally unique path, eliminating the race.

Commit 2 — fix(ui): replace hardcoded colors with semantic Tailwind vars

Files: src/components/Kubernetes/PortForwardList.tsx, src/components/Kubernetes/WorkloadOverview.tsx

Replaced non-adaptive text-gray-* / bg-gray-* classes with text-muted-foreground, bg-muted, border-border — Tailwind CSS vars that correctly invert in dark mode.

Commit 3 — fix(kube): WorkloadOverview loads data; single connect; visible error

Files: src/pages/Kubernetes/KubernetesPage.tsx, tests/unit/KubernetesPage.test.tsx

  • Added case "workloads_overview" in loadResourceData that fetches pods + deployments + statefulsets + daemonsets + jobs + cronjobs via Promise.allSettled in parallel.
  • Added initializedRef guard in loadInitialData to prevent double-connect when selectedClusterId changes.
  • Connection errors now captured and shown as a dismissible banner.

Commit 4 — fix(kube): add namespace to PodInfo; pod actions use pod.namespace

Files: src-tauri/src/commands/kube.rs, src/lib/tauriCommands.ts, src/components/Kubernetes/PodList.tsx, tests/unit/PodList.test.tsx

Added namespace: String to PodInfo Rust struct, extracted from metadata.namespace in parse_pods_json. Added namespace: string to TypeScript PodInfo interface. Updated all 6 action call sites in PodList to use pod.namespace.

Commit 5 — fix(kube): network/config/storage list actions use item.namespace

Files: ServiceList, IngressList, ConfigMapList, SecretList, HPAList, PVCList, ServiceAccountList, RoleList, RoleBindingList, NetworkPolicyList, ResourceQuotaList, LimitRangeList + tests/unit/NamespaceActionFix.test.tsx

12 components fixed. 24 new tests (2 per component).

Commit 6 — fix(kube): workload list actions use item.namespace not filter prop

Files: DeploymentList, StatefulSetList, DaemonSetList, ReplicaSetList, JobList, CronJobList + tests/unit/WorkloadListActions.test.tsx

6 components fixed. 21 new tests.

Testing Needed

  1. Automated: cargo test → 364 pass; npm run test:run → 325 pass; npx tsc --noEmit → 0; npx eslint . --max-warnings 0 → 0; cargo clippy -- -D warnings → 0; cargo fmt --check → clean
  2. Manual — race condition: With a live cluster, rapidly switch between Pods → Deployments → Services → ConfigMaps several times. Data should load reliably every time.
  3. Manual — pod actions: Select "All Namespaces". Open pod action menu → Logs → should fetch without error. Shell/Attach → modals open, exec targets correct namespace. Edit → YAML editor opens.
  4. Manual — overview: Navigate to Workloads → Overview. Cards should show actual pod/deployment/etc. counts.
  5. Manual — error banner: Configure an invalid kubeconfig. Navigate to Kubernetes page. A red banner should appear with the connection error. Clicking Dismiss hides it.
  6. Manual — dark mode: Switch to dark theme. All text in Kubernetes pages (sidebar, tables, status indicators) should be readable with good contrast.