Add red 'stalled' status to admin dashboard banner

[dereuromark]

Summary

The dashboard banner gains a third state so the operator sees a clear "action required" signal when work piles up and nothing is processing it.

state	when
green Queue Running	heartbeat < Queue.dashboardIdleAfter (60s default)
yellow Queue Idle	heartbeat ≥ Queue.dashboardIdleAfter, OR no heartbeat & no backlog
red Queue Stalled	heartbeat ≥ Queue.dashboardStalledAfter (120s default) with a pending backlog and no in-flight job; OR no worker has reported at all and there's a pending backlog or a stuck fetched job

When red, the banner expands with a small diagnostic grid (last activity absolute + relative, workers / servers, pending count) and a one-line cause hint that names the likely fault and the relevant CLI command. See screenshots below.

What red replaces

Two cases were previously ambiguous or hidden:

1. Pending jobs piling up with workers registered but not picking them up — the banner used to say "Queue Idle" indefinitely, regardless of how stale the heartbeat got.
2. Total cron outage — when every queue_processes row aged past Queue.defaultRequeueTimeout, QueueProcessesTable::status() returns empty and the previous template fell through to a muted info notice ("No queue status available"). That's the worst case and is now the loudest red.

Conditions that keep red from false-positiving

• workers == 0 alone does not trigger red. In cron-driven mode that's the normal idle state for a quiet system, so red also requires pendingJobs > 0.
• runningJobs > 0 keeps a busy worker out of red. Heartbeats fire at the top of the worker loop, not during runJob(), so a long-running task (>2 min) with more pending behind it would otherwise look stalled by heartbeat age alone. runningJobs is already computed in the controller (fetched IS NOT NULL AND completed IS NULL).

Config

Two new keys in the example config:

'dashboardIdleAfter' => 60,    // seconds
'dashboardStalledAfter' => 120, // seconds

Defaults are deliberate UI policy — human-perceptible 1-min / 2-min boundaries — not derived from queue mechanics. The existing knobs (workerLifetime, defaultRequeueTimeout, sleeptime) describe exit policy / job-reassignment safety / idle sleep cadence respectively; none of them actually mean "how stale before the dashboard should nag the admin." Installations with unusual cron cadence (e.g. slow exitwhennothingtodo cron) should raise dashboardStalledAfter past the cron interval to avoid false-red between ticks.

Files

• templates/Admin/Queue/index.php — three-state banner, conditional red, no-status path also red on backlog/in-flight, diagnostic grid + cause hint
• templates/layout/queue.php — .status-banner.status-stalled CSS (red gradient + 4 px left accent)
• config/app.example.php — new Queue.dashboardIdleAfter and Queue.dashboardStalledAfter keys with explanation

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 78.34%. Comparing base (06abf17) to head (6e5602a).
⚠️ Report is 1 commits behind head on master.
❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff            @@
##             master     #499   +/-   ##
=========================================
  Coverage     78.34%   78.34%           
  Complexity      981      981           
=========================================
  Files            45       45           
  Lines          3330     3330           
=========================================
  Hits           2609     2609           
  Misses          721      721           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.