# Scheduled Tasks

# Scheduled Tasks

Scheduled tasks run work periodically inside a Curiosity Workspace environment. They are commonly used for:

• periodic ingestion (hourly/daily sync)
• reindexing and maintenance operations
• enrichment jobs (NLP reparse, entity linking, batch similarity)
• reporting and aggregation

# What makes a good scheduled task

• Idempotent: reruns are safe.
• Bounded: does not process unbounded amounts of data without pagination.
• Observable: emits logs/metrics and failure reasons.
• Permission-aware: respects access control where needed.

# Task categories

• Ingestion tasks
  • call connectors or integration logic on a schedule
• Maintenance tasks
  • rebuild indexes, backfills, schema migrations (with care)
• Enrichment tasks
  • reparse fields with updated NLP pipelines
  • compute derived relationships
• Analytics tasks
  • compute dashboards and cached aggregates

# Next steps

• Keep ingestion consistent over time: Ingestion Pipelines
• Run custom logic safely: Custom Endpoints