Monitoring chart installation stuck after first step on 2.15

[izaac]

Setup

• Rancher version: v2.15-3c4f565cf3cc705ed26aba6e4ea2431ef3c954fc-head -- master e464e76
• Rancher UI Extensions: N/A
• Browser type & version: Chrome

Describe the bug

Trying to do a Monitoring Chart app install the UI is stull loading after the first step (after clicking next)

To Reproduce

Only on Downstream clusters

• Install a head Rancher build or latest 2.15 alpha.
• Go to Apps and select to install the Monitoring chart.
• Select version, click next on the first step's Next button

Result

UI Stuck in loading.

Expected Result
Monitoring install continue properly to second step and not get stuck in loading.

Screenshots

rec-20260521-091839.mp4

It silently fails without errors in the Browser console.

[github-actions]

🎯 Agentic Issue Triage

Summary: The Monitoring chart installation flow gets stuck in a loading state after clicking "Next" on the first step in Rancher v2.15. No errors appear in the browser console, making this a silent failure.

📋 Issue Analysis

• Type: Bug (confirmed by issue type field)
• Milestone: v2.15.0
• Affected Area: Apps/Charts — Monitoring chart installation wizard
• Severity: High — the Monitoring chart is a core feature and this blocks installation entirely
• Regression: Likely a regression introduced in the master/2.15 line (specific commit e464e76)
• User Impact: Users cannot install the Monitoring chart through the UI on v2.15 builds

🔍 Debugging Strategies

1. Network tab inspection: Open DevTools → Network tab, reproduce the issue, look for failed or hanging API requests after clicking "Next". The step transition likely triggers a fetch (e.g., chart values, schema, or app version details) that hangs.
2. Vue DevTools: Inspect the loading state of the AsyncButton or the stepper component — check which reactive property is stuck as true.
3. Check chart-specific queries: The first-to-second step transition for chart installs typically fetches questions.yaml, chart values, and optionally CRDs. Check if any of these calls are pending indefinitely.
4. Console errors with verbose logging: Try enabling verbose logging or check window.__vue_devtools_global_hook__ for unhandled promise rejections.
5. Git bisect: Narrow the regression between the last known good commit and e464e76 on master to identify the offending change.

🧭 Relevant Code Areas

• shell/pages/c/_cluster/apps/charts/install.vue — chart install wizard page
• shell/components/chart/ — chart step components
• shell/store/catalog.js / shell/store/catalog.ts — catalog/chart data fetching
• Look for any recent changes to the chart install stepper or the API calls triggered on step transition

🔗 Possible Reproduction Steps

1. Deploy a Rancher v2.15 head or latest 2.15 alpha build
2. Navigate to a downstream cluster → Apps → Charts
3. Select the Monitoring chart
4. Choose a version and click Next on Step 1
5. Observe the UI enters a loading state and never proceeds to Step 2
6. Confirm no errors appear in the browser console

✅ Suggested Sub-tasks

• Reproduce the issue on a fresh v2.15 head build
• Identify the exact API call or state mutation that causes the hang (Network tab / Vue DevTools)
• Check for recent changes to the chart install wizard or catalog store around e464e76
• Verify whether other charts (e.g., Longhorn, NeuVector) are also affected or if this is Monitoring-specific
• Fix the root cause and add a regression test or unit test for the step transition logic
• Verify fix on v2.15 head

Generated by Agentic Triage for issue #17765 · ● 171.7K · ◷

[izaac]

Additional Observations:

the api call to /k8s/clusters/c-m-npj5lq7h/v1/secrets/cattle-monitoring-system?pagesize=100000&exclude=metadata.managedFields is stuck in (pending) status.

The UI loads step 2 only after that errors with net::ERR_NETWORK_CHANGED which could be quick or take a few minutes.

[github-actions]

🤖 Issue grooming questions

The problem description, reproduction steps, and setup details are clear. Two items are needed before this can be triaged for development.

1. Missing: Acceptance Criteria

Please add a concrete, testable checklist defining what "fixed" looks like. For example:

• Clicking Next on the first step of the Monitoring chart installation on a downstream cluster proceeds to the second step without getting stuck in a loading state.
• No indefinite loading spinner appears during a Monitoring chart installation on a downstream cluster.

2. Missing: Technical Details

The issue notes "it silently fails without errors in the browser console." To help pinpoint the root cause, could you check:

• Are there any failed or unexpected network requests visible in the browser Network tab while the UI is stuck (e.g. a request returning a 4xx/5xx, or a request that never completes)?
• Does the loading state correspond to a specific API call or component initialisation (e.g. a values fetch, schema fetch, or chart step transition)?
• Is this reproducible only on v2.15/master, or also on an earlier version?

This comment was generated automatically to help gather the information needed to groom this issue.

Generated by Daily Issue Grooming · ● 870.5K · ◷