admin/mcp-nextcloud
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: Increase placeholder staleness threshold to 5x scan interval

Browse Source 
- Changed from 2x (120s) to 5x (300s) scan interval
- Large PDFs take 3-4 minutes to process, need longer threshold
- Prevents premature requeuing of in-flight documents
This commit is contained in:
Chris Coutinho
2025-11-20 15:36:49 +01:00
parent 47f0b3db9a
commit ec2c274cd9
2 changed files with 17 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files
nextcloud_mcp_server/observability/logging_config.py
+11 -8
View File
@@ -37,14 +37,17 @@ class HealthCheckFilter(logging.Filter):
""" """
# Check if the log message contains health check endpoints # Check if the log message contains health check endpoints
message = record.getMessage() message = record.getMessage()
return not any( return (
endpoint in message not any(
for endpoint in [ endpoint in message
"/health/live", for endpoint in [
"/health/ready", "/health/live",
"/metrics", "/health/ready",
"/app/vector-sync/status", "/metrics",
] "/app/vector-sync/status",
]
)
or "OpenCV not installed. Disabling OCR" in message
) )
nextcloud_mcp_server/vector/scanner.py
+6 -4
View File
@@ -266,10 +266,11 @@ async def scan_user_documents(
needs_indexing = True needs_indexing = True
elif existing_metadata.get("is_placeholder", False): elif existing_metadata.get("is_placeholder", False):
# Placeholder exists - check if it's stale (processing may have failed) # Placeholder exists - check if it's stale (processing may have failed)
# Only requeue if placeholder is older than 2x scan interval # Only requeue if placeholder is older than 5x scan interval
# (Large PDFs can take 3-4 minutes to process)
queued_at = existing_metadata.get("queued_at", 0) queued_at = existing_metadata.get("queued_at", 0)
placeholder_age = time.time() - queued_at placeholder_age = time.time() - queued_at
stale_threshold = get_settings().vector_sync_scan_interval * 2 stale_threshold = get_settings().vector_sync_scan_interval * 5
if placeholder_age > stale_threshold: if placeholder_age > stale_threshold:
logger.debug( logger.debug(
f"Found stale placeholder for note {doc_id} " f"Found stale placeholder for note {doc_id} "
@@ -460,10 +461,11 @@ async def scan_user_documents(
needs_indexing = True needs_indexing = True
elif existing_metadata.get("is_placeholder", False): elif existing_metadata.get("is_placeholder", False):
# Placeholder exists - check if it's stale (processing may have failed) # Placeholder exists - check if it's stale (processing may have failed)
# Only requeue if placeholder is older than 2x scan interval # Only requeue if placeholder is older than 5x scan interval
# (Large PDFs can take 3-4 minutes to process)
queued_at = existing_metadata.get("queued_at", 0) queued_at = existing_metadata.get("queued_at", 0)
placeholder_age = time.time() - queued_at placeholder_age = time.time() - queued_at
stale_threshold = get_settings().vector_sync_scan_interval * 2 stale_threshold = get_settings().vector_sync_scan_interval * 5
if placeholder_age > stale_threshold: if placeholder_age > stale_threshold:
logger.debug( logger.debug(
f"Found stale placeholder for file {file_path} (ID: {file_id}) " f"Found stale placeholder for file {file_path} (ID: {file_id}) "