Tutorial: Building with CleaveDB
Scheduled Tasks & Triggers — Overview
Traditional database stacks require an assortment of external tooling—such as system cron daemons, Celery queues, Airflow workflows, and Kafka/Debezium CDC pipelines—just to handle periodic cleanup, aggregated reports, and event propagation. CleaveDB eliminates this architectural sprawl with native, kernel-level automation.
The Three Pillars of Automation
1. Repeating Tasks (EVERY)
Schedule recurring queries (interval cleanups, nightly health checks, periodic distill rollups) using human-readable cron expressions like EVERY 5 MINUTES DO (...).
2. Database Triggers (ON ... RUN)
Execute reactive CleaveQL operations when documents are inserted or modified, with dynamic key interpolation ($gid, $amount).
3. Webhooks for CDC
Publish real-time Change Data Capture payloads to external HTTP endpoints on writes, updates, and deletes without configuring message queues.
Execution Safety & Isolation
Every automated routine in CleaveDB is subject to the database engine's safety guarantees:
- Recursion Depth Protection: Triggers maintain an execution context counter (
trigger_depth) capped at 5. Any circular trigger relationship is halted before causing stack overflow panics. - Transactional Atomicity: When triggered within an active
BEGIN ... COMMITblock, trigger errors immediately roll back uncommitted parent mutations. - Audit Attribution: Scheduled tasks execute with authenticated session context and append execution records into the system audit log.
Explore the Automation Topics
Follow the sections below to configure automated routines in CleaveDB:
- EVERY — repeating tasks: configure recurring cron intervals for maintenance, schema migrations, and analytics.
- ON ... RUN — triggers: create reactive event listeners that fire on document writes with dynamic field interpolation.
- Webhooks for CDC: stream real-time JSON event mutations directly to HTTP endpoints without message brokers.