aksd: MetricsTab & MetricsCard: Separate presentation from logic; Tests, refactors, documentation

[tejhan]

Description

This PR splits the MetricsTab and MetricsCard into modular components & extracts functionality into hooks following the presentation/logic separation pattern outlined in #97 & #120 .

Changes Made

• Created useNamespaceLabels, useDeployments, usePods, usePrometheusMetrics, useCardMetrics hooks
• Created DeploymentSelector, EmptyStateCard, MetricsSummaryBar, MetricsChart, MetricsChartsGrid, MetricsLoadingSkeleton, PodDetailsTable, MetricStatCard components
• Created types ChartDataPoint, ResponseTimeDataPoint, MetricSummary, MemoryUnit & functions pickMemoryUnit, convertBytesToUnit, formatMemoryBrief
• Prometheus utils moved to shared utils/prometheus/ location
• Implemented more explicit cleanup functions
• Fixed rendering issues during loading state
• Full Tsdoc documentation for relevant elements
• Implemented caching for all metrics
• UI fixes & alignment

Rebase Changes

In order to preserve the translation preparation that was done with useTranslation, I've moved the useTranslation invokations to match the same strings but in their new file locations across the components/hooks.

All Metrics related content is now in the shared Metrics folder.

Translations, new constants & UI changes have all been included in rebase.

Type of changes

• [ x] Bug fix (non-breaking change which fixes an issue)
• [x ] Code refactoring

Related Issues

Closes #97, #120

• plugins/aks-desktop: Separate presentation from logic plugins/aks-desktop/src/components/MetricsTab/MetricsTab.tsx MetricsTab #97
• plugins/aks-desktop: Separate presentation from logic plugins/aks-desktop/src/components/Metrics/MetricsCard.tsx MetricsCard #120

Testing

For automated testing, you can either run npm run plugin:test for the full suite including these new ones. If you want to test these suites alone, you can run cd plugins/aks-desktop, and then npx vitest run src/components/Metrics.

To manually test the components / data are rendering properly,

1. Create/use an AKS cluster with Prometheus monitoring enabled.
2. Create an Aks managed project
3. Deploy an application to that project
4. Observe metrics tab populate.

Test Cases

Describe the test cases that were run:

1. Full test suite for all hooks & utility functions.
2. Manual testing via creating & utilizing existing projects & observing if Metrics Tab properly loads components initially, and updates state correctly after time.

Documentation Updates

• [x ] Code comments updated

[Copilot]

Pull request overview

Refactors the AKS Desktop plugin’s MetricsTab to follow a presentation/logic separation approach (per #97), while also relocating Prometheus helper utilities into a shared utils/prometheus location for reuse across tabs.

Changes:

• Split MetricsTab into dedicated hooks (useDeployments, usePods, useNamespaceLabels, usePrometheusMetrics) and presentational components (charts grid, summary bar, selectors, loading/empty states, pod table).
• Introduced shared MetricsTab utility types/helpers for memory unit selection and formatting.
• Updated other features (ScalingTab, MetricsCard) to import Prometheus helpers from the new shared utils location.

Reviewed changes

Copilot reviewed 17 out of 17 changed files in this pull request and generated 9 comments.

Show a summary per file

File	Description
plugins/aks-desktop/src/utils/prometheus/queryPrometheus.tsx	Adjusts import path and adds TSDoc for shared Prometheus querying utility.
plugins/aks-desktop/src/utils/prometheus/getPrometheusEndpoint.tsx	Adjusts import paths and adds TSDoc for shared endpoint discovery utility.
plugins/aks-desktop/src/components/ScalingTab/ScalingTab.tsx	Updates Prometheus helper imports to shared utils location.
plugins/aks-desktop/src/components/MetricsTab/utils.ts	Adds shared MetricsTab types and memory formatting/unit conversion helpers.
plugins/aks-desktop/src/components/MetricsTab/hooks/usePrometheusMetrics.ts	New hook to fetch/poll Prometheus and transform results for charts/summary.
plugins/aks-desktop/src/components/MetricsTab/hooks/usePods.ts	New hook to list pods for the selected deployment and derive status/total pod count.
plugins/aks-desktop/src/components/MetricsTab/hooks/useNamespaceLabels.ts	New hook to read subscription/resource-group labels from the namespace metadata.
plugins/aks-desktop/src/components/MetricsTab/hooks/useDeployments.ts	New hook to list deployments and manage initial selection/loading/error.
plugins/aks-desktop/src/components/MetricsTab/components/PodDetailsTable.tsx	Extracted pod table into a dedicated presentational component.
plugins/aks-desktop/src/components/MetricsTab/components/MetricsSummaryBar.tsx	Extracted metrics summary bar into a dedicated component.
plugins/aks-desktop/src/components/MetricsTab/components/MetricsLoadingSkeleton.tsx	Extracted loading skeleton UI into a dedicated component.
plugins/aks-desktop/src/components/MetricsTab/components/MetricsChartsGrid.tsx	New component composing the set of metric charts into a grid layout.
plugins/aks-desktop/src/components/MetricsTab/components/MetricsChart.tsx	New reusable chart wrapper component for Recharts line charts.
plugins/aks-desktop/src/components/MetricsTab/components/EmptyStateCard.tsx	Extracted empty-state UI into a reusable component.
plugins/aks-desktop/src/components/MetricsTab/components/DeploymentSelector.tsx	Extracted deployment selector dropdown into its own component.
plugins/aks-desktop/src/components/MetricsTab/MetricsTab.tsx	Simplifies MetricsTab to composition of hooks + presentational components.
plugins/aks-desktop/src/components/Metrics/MetricsCard.tsx	Updates Prometheus helper imports to shared utils location.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[skoeva]

could you write unit tests for this? I think TDD is the move and would be very helpful to for these component refactors moving forward

[tejhan]

could you write unit tests for this? I think TDD is the move and would be very helpful to for these component refactors moving forward

Yep, currently working on these.

[Copilot]

Pull request overview

Copilot reviewed 25 out of 25 changed files in this pull request and generated 5 comments.

Comments suppressed due to low confidence (1)

plugins/aks-desktop/src/components/MetricsTab/MetricsTab.tsx:96

• The error state mixes error (deployments) and metricsError (Prometheus) but the UI only renders {error}. When metricsError is set and error is null, the alert body will be blank. Also the AlertTitle string is hard-coded and bypasses i18n. Prefer rendering t('Metrics Unavailable') and show error ?? metricsError (and consider surfacing metricsError even when deployments exist so Prometheus failures are visible).

  if ((error || metricsError) && deployments.length === 0) {
    return (
      <Box p={3}>
        <Alert severity="warning">
          <AlertTitle>Metrics Unavailable</AlertTitle>
          <Typography variant="body2">{error}</Typography>
        </Alert>

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[illume]

@skoeva can you please check the changes in here relating to ScalingTab? I don't know if they conflict with yours?

[illume]

I read through all this and it's looking good IMHO.

• Can you please describe how to manually test this in the PR description?
• I think the commit message subject could have the context part added (maybe just MetricsTab) and a longer description of the changes.
• Can you please look at the open review comments?

[Copilot]

Pull request overview

Copilot reviewed 25 out of 25 changed files in this pull request and generated 5 comments.

Comments suppressed due to low confidence (1)

plugins/aks-desktop/src/components/MetricsTab/MetricsTab.tsx:96

• The error banner is gated on deployments.length === 0, and within the banner it renders {error} even when the failure came from metricsError. This can result in a warning with an empty message and also hides Prometheus fetch errors once deployments are loaded. Consider (1) rendering error ?? metricsError and (2) surfacing metricsError even when deployments exist (e.g., inline alert above charts). Also Metrics Unavailable is currently hard-coded instead of translated via t(...).

  if ((error || metricsError) && deployments.length === 0) {
    return (
      <Box p={3}>
        <Alert severity="warning">
          <AlertTitle>Metrics Unavailable</AlertTitle>
          <Typography variant="body2">{error}</Typography>
        </Alert>

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[skoeva]

@skoeva can you please check the changes in here relating to ScalingTab? I don't know if they conflict with yours?

@illume they only involve import statements from the MetricsTab code, there shouldn't be any conflict

[Copilot]

Pull request overview

Copilot reviewed 30 out of 30 changed files in this pull request and generated 4 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 30 out of 30 changed files in this pull request and generated 5 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 30 out of 47 changed files in this pull request and generated 3 comments.

Comments suppressed due to low confidence (1)

plugins/aks-desktop/src/utils/prometheus/getPrometheusEndpoint.tsx:14

• The TSDoc describes only a successful return value, but this function throws in multiple failure cases (e.g., when Azure Monitor Metrics isn't enabled, no matching rule group is found, JSON parse fails, etc.). Please document the thrown error behavior (e.g., add an @throws tag and/or clarify in @returns that errors are thrown).

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 29 out of 29 changed files in this pull request and generated 3 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 29 out of 29 changed files in this pull request and generated no new comments.

Comments suppressed due to low confidence (1)

plugins/aks-desktop/src/utils/prometheus/queryPrometheus.tsx:38

• queryPrometheus runs az account get-access-token on every PromQL query. MetricsTab issues multiple queries per refresh (and Scaling now also uses this util), so this can spawn many az processes every 30s and significantly slow the UI / hit rate limits. Consider acquiring the token once per refresh (or caching tokens per subscription until expiry) and reusing it across queries, or make queryPrometheus accept an already-acquired token.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[illume]

Great work, thanks! 🎉

These changes were done quite some time ago.