fix(selfhost): gate maintenance admission on genuine live pressure, not raw pending count

[JSONbored]

Summary

• evaluateMaintenanceAdmission's live_pending_high/live_job_age_high checks gated on livePendingCount/oldestLivePendingAgeMs — a raw pending/processing count and created-at age that include work deliberately deferred to the future (e.g. agent-regate-pr's normal, expected, staggered per-PR backlog, index.ts:24-29). On edge-nl-01 this let a routine backlog (not real load: host load was ~0.01/core, zero foreground jobs runnable for a sustained 100s sample) starve the maintenance lane for hours — 69% of admission denials in an 18-minute window were live_pending_high, and the maintenance_pending_high drain escape built specifically to prevent lane self-pinning never got a chance to fire because this check denies first.
• Switches both checks to the runnable-now signal (liveRunnableNowCount/oldestLiveRunnableAgeMs), and broadens that signal's SQL in both queue backends (sqlite + pg) to also count actively processing foreground jobs as genuine current pressure, not just due-and-unclaimed pending ones — a processing job is real, in-flight resource use and arguably more legitimate pressure than a merely-due-but-waiting one.
• No new config surface: the existing global MAINTENANCE_ADMISSION_MAX_LIVE_PENDING/MAINTENANCE_ADMISSION_MAX_LIVE_AGE_MS env knobs are unchanged — they now simply measure what their doc comments always said they measured.

Closes #4669

Scope

• The PR title follows type(scope): short summary Conventional Commit format, for example fix(api): restore profile access checks.
• This PR is focused and does not mix unrelated backend, UI, MCP, docs, dependency, and deploy changes.
• This follows CONTRIBUTING.md and does not reintroduce GitHub Pages, VitePress, site/, or CNAME.
• I linked a currently open issue this PR resolves (e.g. Closes #123) — a linked open issue is required for every contributor PR.

Validation

• git diff --check
• npm run actionlint
• npm run typecheck
• npm run test:coverage locally; codecov/patch requires ≥99% coverage of the lines AND branches you changed (aim for 100% on your diff so CI variance does not fail near the threshold). Global coverage is a non-blocking trend with a loose 90% backstop, not the gate.
• npm run test:workers
• npm run build:mcp
• npm run test:mcp-pack
• npm run ui:openapi:check
• npm run ui:lint
• npm run ui:typecheck
• npm run ui:build
• npm audit --audit-level=moderate
• New or changed behavior has unit/integration tests for new branches, fallback paths, and sanitizer boundaries

If any required check was skipped, explain why:

• This is a backend-only change to src/selfhost/{maintenance-admission,sqlite-queue,pg-queue}.ts with no UI, MCP, OpenAPI, wrangler binding, or workflow surface, so actionlint/test:workers/build:mcp/test:mcp-pack/ui:* are not applicable. test:coverage was run targeted (vitest run test/unit/selfhost-{maintenance-admission,sqlite-queue,pg-queue}.test.ts --coverage, 330 tests, all green) rather than the full unsharded suite; maintenance-admission.ts's only reported uncovered line (106) is a pre-existing JSDoc comment line outside this diff, unrelated to the change.

Safety

• No secrets, wallet details, hotkeys, coldkeys, user PATs, private keys, raw trust scores, private rankings, or private maintainer evidence are exposed.
• Public GitHub text stays sanitized, low-noise, and does not imply compensation guarantees or optimization tactics.
• Auth, cookie, CORS, GitHub App, Cloudflare, or session changes include negative-path tests. (N/A — no auth/session surface touched)
• API/OpenAPI/MCP behavior is updated and tested where needed. (N/A — internal queue admission logic only, no external API surface)
• UI changes use live API data or real empty/error/loading states, not production mock/demo fallbacks. (N/A — no UI changes)
• Visible UI changes include a UI Evidence section below with JPG/JPEG or PNG screenshots arranged as organized, captioned, clickable thumbnails. (N/A — no visible UI change)
• Public docs/changelogs are updated where needed; changelogs are only edited for release-prep PRs. (N/A)

UI Evidence

N/A — no UI/frontend/docs-visible change; this is an internal self-host queue admission fix.

Notes

• Live evidence (host load, queue metrics, denial-reason breakdown) that motivated this fix was gathered directly from edge-nl-01's /metrics endpoint and its Postgres _selfhost_jobs table during the investigation; details are in fix(selfhost): maintenance-admission gates on raw live-pending count, not runnable-now, starving the maintenance lane #4669.

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.13%. Comparing base (6dcd997) to head (6332f3d).
⚠️ Report is 6 commits behind head on main.
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #4674   +/-   ##
=======================================
  Coverage   94.13%   94.13%           
=======================================
  Files         434      434           
  Lines       38541    38541           
  Branches    14054    14054           
=======================================
  Hits        36281    36281           
  Misses       1601     1601           
  Partials      659      659           

Files with missing lines	Coverage Δ
src/selfhost/maintenance-admission.ts	100.00% <100.00%> (ø)
src/selfhost/sqlite-queue.ts	99.60% <ø> (ø)

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.