my prefect server setup prefect-metrics.waow.tech
python orchestration
0
fork

Configure Feed

Select the types of activity you want to include in your feed.

fix: snapshot DuckDB before reading to bypass flock

same pattern as the hub frontend — copy the DB file to /tmp before
connecting, since Grafana holds an exclusive flock on the source.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

zzstoatzz 1e96c804 a883bb5f

+5 -1
+5 -1
flows/curate.py
··· 1 1 import os 2 + import shutil 2 3 from datetime import datetime, timezone 3 4 from pathlib import Path 4 5 ··· 45 46 @task 46 47 def load_items(db_path: str) -> str: 47 48 """Read scored items from hub_action_items, format as text for the LLM.""" 48 - db = duckdb.connect(db_path, read_only=True) 49 + # snapshot to bypass exclusive flock (same pattern as hub frontend) 50 + snap = "/tmp/curate_analytics_snapshot.duckdb" 51 + shutil.copy2(db_path, snap) 52 + db = duckdb.connect(snap, read_only=True) 49 53 rows = db.execute( 50 54 "SELECT source, repo, identifier, kind, title, url, " 51 55 "author, labels, importance_score, updated "