Replace Static "Not Installed" Message with Installation Diagnostics Panel

[atharrva01]

Summary

Right now, if any pod check fails, we just show "Inspektor Gadget is not installed" with a docs link. This is unhelpful because pods could be crashing, stuck in CrashLoopBackOff, or partially deployed, but the UI just says "not installed." Users have to jump to kubectl to figure out what's actually wrong.

What I Changed

Added a basic diagnostics panel that checks:

• Namespace exists (gadget)
• DaemonSet status (desired vs ready replicas)
• Pod health (crash loops, image pull errors, pending)
• Node coverage (are gadget pods on all schedulable nodes?)

Each check shows pass/warn/fail with details you can expand. Way more useful than a static error message.

Why This Helps

Instead of guessing, users can now see if IG is actually missing vs just misconfigured or failing to start.

Testing

Built locally, ran tsc --noEmit and npm run lint. All good.

[atharrva01]

hi @ashu8912 this PR replaces the static “Inspektor Gadget is not installed” message with an installation diagnostics panel that performs checks (namespace, DaemonSet, pod health, node coverage) and surfaces actionable troubleshooting information directly in the Headlamp UI.

[Copilot]

Pull request overview

This PR replaces the generic “Inspektor Gadget is not installed” screen with an installation diagnostics panel that surfaces specific cluster/deployment health checks (namespace, DaemonSet, pods, node coverage) to help users understand what’s actually wrong.

Changes:

• Replaced the static “not installed” message with a new InstallationStatus diagnostics UI.
• Added Kubernetes checks for namespace existence, DaemonSet readiness, pod health, and node coverage.
• Added “Copy Diagnostics” and “Re-run Checks” actions to help share troubleshooting context.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 6 comments.

File	Description
src/common/NotFound/index.tsx	Swaps the old static not-installed message for the new diagnostics component.
src/common/InstallationStatus/index.tsx	Implements the diagnostics panel UI and wiring to K8s useList data sources.
src/common/InstallationStatus/checks.ts	Implements the underlying diagnostic checks and formatting logic.

---

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

[atharrva01]

hey @ashu8912 looking at the co-pilot's review, I have fixed the DaemonSet selection (.find() with k8s-app=gadget instead of [0]), scoped Pod.useList() to IG_NAMESPACE which also fixes the namespace leak in checkPodHealth and checkNodeCoverage, added .catch() on clipboard, and tightened StatusIcon to CheckResult['status'].

Skipped prop-hoisting for now, this only renders when IG isn't installed so not a hot path, happy to do it as a follow-up.

[Copilot]

Pull request overview

This PR replaces the generic “Inspektor Gadget is not installed” message with an Installation Diagnostics panel that surfaces likely deployment/health issues (namespace/DaemonSet/pods/node coverage) directly in the UI, reducing the need to switch to kubectl.

Changes:

• Replaced the IGNotFound static message with the new InstallationStatus diagnostics view.
• Added a new diagnostics UI (overall banner + expandable per-check results + copy-to-clipboard + retry).
• Implemented installation/health check functions for namespace, DaemonSet readiness, pod health, and node coverage.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 4 comments.

File	Description
src/common/NotFound/index.tsx	Switches “not installed” fallback to render InstallationStatus.
src/common/InstallationStatus/index.tsx	Adds the diagnostics panel UI, check rendering, copy diagnostics, and retry action.
src/common/InstallationStatus/checks.ts	Adds the underlying diagnostics checks and formatting utilities.

---

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

[atharrva01]

hi @ashu8912 , @illume! addressed the remaining copilot suggestions from the second review:

• Clipboard guard : added an availability check before calling the API, shows an error snackbar if it fails
• Recalculate button : renamed from "Re-run Checks" since the hooks are live watchers, not fresh fetches
• Cordoned nodes : checkNodeCoverage now catches spec.unschedulable === true too, so kubectl cordon nodes are properly excluded
• Pod crash reason : tightened .find() to only match the same states as isCrashing, so ContainerCreating can't sneak in as a crash cause

[Copilot]

Pull request overview

Copilot reviewed 6 out of 6 changed files in this pull request and generated 6 comments.

---

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

[Copilot]

Pull request overview

Copilot reviewed 6 out of 6 changed files in this pull request and generated 4 comments.

---

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

[illume]

Thanks @atharrva01

Much appreciated 🎉

Would you mind having a look at the copilot suggestions?

[atharrva01]

@illume , i've addressed the co-pilot's suggestion....

[illume]

Thanks! 🎉