fix: record overrun candles as SKIPPED_OVERRUN, not data-unavailable (#169)

[mohammadp1001]

Fixes #169.

A candle missed because our own processing overran was filed as SKIPPED_DATA_UNAVAILABLE - "we could not get market data". The data was fine. The previous session was still running.

That matters because SKIPPED_DATA_UNAVAILABLE is the signal you check when you suspect the Alpaca feed is broken. Overrun candles polluted it, so a run showing eight "data unavailable" rows sent you off to check the feed when the real story was an investigation running longer than its candle. Two different problems, two different fixes, previously indistinguishable in the memory bank and in the session report.

What changed

• SKIPPED_OVERRUN added as a Session status, used for the market-open branch of _handle_overrun_candles.
• SKIPPED_MARKET_CLOSED unchanged - that branch was already correct.
• SKIPPED_DATA_UNAVAILABLE keeps its meaning and its other uses (a failed investigation still returns it).
• Both statuses were already outside the FR-018 completed count, so run length is unaffected.

Spec

data-model.md gains the new value, plus a table saying what each status actually means - the previous one-line list did not explain why you would see one rather than another.

605 passed in 142.42s (0:02:22)
Required test coverage of 100% reached. Total coverage: 100.00%
ruff check alphoryn/ tests/ -> All checks passed!

[kilo-code-bot]

Code Review Summary

Status: No Issues Found | Recommendation: Merge

Files Reviewed (4 files)

• alphoryn/memory/schema.py
• alphoryn/scheduler/scheduler.py
• specs/001-etf-paper-trading-agent/data-model.md
• tests/unit/test_scheduler.py

---

_{Reviewed by step-3.7-flash · Input: 67.5K · Output: 14.6K · Cached: 1M}