fix(ingest): address review round 1 (reclaim queue + tests)

- Register the periodic stalled-job reclaim on a dedicated ingest-maintenance
  queue that every worker drains (any --tier), so reclaim still fires when the
  fast fleet is scaled to zero and only ocr workers run. procrastinate's
  periodic-defer dedup keeps it single-run across drainers.
- escalation: mark `unsupported`/`forced` reason labels as reserved (not raised).
- processor: note that options/progress_callback are intentionally not threaded
  through _parse_pdf_tier yet (symmetric with the inline path).
- tests: assert TieredEscalationStrategy backoff progression (4/8/16/…/300s);
  cover get_ingest_pending per-queue aggregation + the legacy job_counts
  fallback; add an external-path zero-page no-escalation case; use the canonical
  INGEST_QUEUE_FAST instead of the back-compat alias.

Deck #323.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
Chris Coutinho
2026-06-13 13:34:03 +02:00
co-authored by Claude Opus 4.8
parent 9676bb3106
commit 35f8204a16
8 changed files with 86 additions and 11 deletions
@@ -5,6 +5,8 @@ and the procrastinate TieredEscalationStrategy that turns a raised exception
into a queue-hop / same-tier retry / give-up decision.
"""
from datetime import datetime, timezone
import httpx
import pytest
from procrastinate.jobs import Job
@@ -87,6 +89,20 @@ class TestTieredEscalationStrategy:
assert decision.queue is None
assert decision.retry_at is not None
def test_transient_backoff_progression(self):
# min(4 * 2**(attempts-1), 300): 4, 8, 16, ... capped at 300s.
strat = self._strategy(max_transient=100)
for attempts, expected in [(1, 4), (2, 8), (3, 16), (4, 32), (20, 300)]:
decision = strat.get_retry_decision(
exception=httpx.ConnectError("x"), job=_job(attempts=attempts)
)
assert decision is not None and decision.retry_at is not None
delta = (decision.retry_at - datetime.now(timezone.utc)).total_seconds()
# retry_at = now + wait; allow a small window for execution time.
assert expected - 2 <= delta <= expected + 1, (
f"attempts={attempts}: delta={delta:.2f}s, expected≈{expected}s"
)
def test_transient_gives_up_over_cap(self):
decision = self._strategy(max_transient=5).get_retry_decision(
exception=httpx.ConnectError("refused"), job=_job(attempts=5)