ETL workflows
Pipeline catalogue
Each workflow models a pattern I use in Oracle-centric platforms: change capture, conformed dimensions, micro-batch aggregation and retention enforcement. Run histories and metrics are generated demo data, not readings from a live database.
Incremental change capture from a mock order-entry schema into a partitioned billing fact table, with currency normalisation and late-arriving-record handling.
- Schedule
- Every 15 min
- SLA
- 20 min
- Healthy runs
- 66.7%
- Last reject rate
- 0.08%
Slowly changing dimension (Type 2) build for a synthetic customer master, including tokenisation of sensitive attributes before load.
- Schedule
- Hourly
- SLA
- 45 min
- Healthy runs
- 66.7%
- Last reject rate
- 1.94%
Micro-batch aggregation of fabricated device telemetry into hourly rollups with idempotent re-processing on replay.
- Schedule
- Every 5 min
- SLA
- 10 min
- Healthy runs
- 100%
- Last reject rate
- 0%
Retention enforcement job that moves aged partitions to a compressed archive tablespace and purges expired synthetic records.
- Schedule
- Nightly 02:00
- SLA
- 90 min
- Healthy runs
- 100%
- Last reject rate
- 0%