[Starosdev]

Summary

WP-E (server operations) for the 2.0.0 release. On-demand collector runs can no longer overlap and now run every bundled collector, and /api/summary reports when each device last reported and whether it is stale, with optional auto-archiving of devices that stop reporting.

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation update
• Refactoring (no functional changes)
• CI/CD or infrastructure change

Related Issues

Related to #965. Addresses #533.

Changes Made

• POST /api/collectors/run: a request while a run is active returns 409 (breaking footer on 39bd9bc). The handler iterates the shared omnibus.CollectorSpecs, so btrfs, filesystem, and performance collectors also run; binaries missing from PATH are skipped. Route group unchanged; admin scope comes with the API token work.
• /api/summary (plain and paginated): per device last_seen (omitted when the device never reported) and stale (omitted when last-seen data could not be loaded), computed with the same rule the missed-ping monitor uses (new models.Device.MissedPingTimeout and IsStale, now shared by both). Archived and never-seen devices are never stale. Works whether or not missed-ping notifications are enabled. A 0 or negative global timeout falls back to the 60-minute default, as in the monitor. Muted devices still report stale. Last-seen data is cached for 30 seconds; a query failure is logged and leaves both fields unset rather than failing the summary.
• web.devices.auto_archive_stale_days (config file, default 0 = off): the monitor tick on the leader archives devices silent for longer than max(days, the device's missed-ping timeout) through the existing archive path, removes them from MQTT, and sends one notification per tick listing them. Never-seen devices are skipped. Auto-archived devices stay archived until unarchived manually. A startup warning is logged when the setting is shorter than the global missed-ping timeout.
• docs/openapi.yaml (summary fields) and example.scrutiny.yaml (new setting).

Testing

• I have tested this locally
• I have added/updated tests that prove my fix/feature works
• All existing tests pass

go build ./..., go vet ./..., go test ./..., go test -race on web, web/handler, models; golangci-lint run --new-from-rev=origin/beta (v2.14.0, linux/amd64) 0 issues. Tests: concurrent run returns 409, all available collectors invoked, stale rule with per-device override, summary fields, auto-archive.

Checklist

• My code follows the project's code style
• I have performed a self-review of my code
• I have commented my code where necessary (particularly complex areas)
• I have updated the documentation if needed
• My changes generate no new warnings

Notes

• The manual-run lock is in-process and does not serialize with omnibus scheduled runs, which are separate child processes. Same as before this PR.
• A persistent auto_archived flag (so a returning device could be unarchived automatically) needs a migration; deferred to the migration work.
• The auto-archive setting is config-file only; a settings-UI version needs a seeded settings row (migration work).
• A device that stops being registered after a failed smartctl --info (fix(collector): skip unidentified devices, add exclude list and bounded parallel collection #969) goes stale only if it already has history; that path is not covered by a dedicated test.
• Frontend follow-up: stale badge and last-seen on dashboard and detail, stale filter.

Review: not required (server change on the beta integration branch; an independent read-only review is being run per release coordination)