feat(ops): device/collection health evaluator and DTOs #100

Parent

Epic: #99
Design: doc/ops-monitor.md (§4–5, §8, Issue A)

Summary

Add a pure, testable ActiveSync ops health model under src/:

• Horde\ActiveSync\Ops\HealthStatus (ok / warn / critical)
• SignalCode reason codes
• CollectionHealth, DeviceHealth, FleetSummary
• HealthEvaluator (heuristics)
• DeviceHealthFactory (Device + SyncCache → DeviceHealth)

No CLI/UI in this issue.

Heuristics (v1)

• Heartbeat: lasthbsyncstarted vs lastsyncendnormal → hb_in_flight / hb_stuck / hb_missing_end
• foldersyncrequired counter → warn / critical at guard ceiling
• Activity window from SyncCache timestamp
• Blocked / wipe-pending as signals
• Tunables as method args / options object (CLI flags later)

Acceptance

• Pure unit tests for every reason code and severity aggregation (no DB)
• Reason codes documented in doc/ops-monitor.md
• PSR-4 / composer autoload for src/ Ops types if needed
• No protocol behaviour change

Notes

Honest limit: true mid-Sync “stuck request” needs later request touchpoints (follow-up). v1 is state-derived only.