Performance Tuning

Parallelism

runtime:
  parallelism: 8
  partition_strategy: hash
  partition_keys: [user_id]
`

## Caching

```yaml
runtime:
  cache:
    enabled: true
    ttl: 3600
    strategy: incremental
`

## Memory Management

```yaml
runtime:
  memory:
    driver: 8GB
    executor: 16GB
    overhead: 1GB
`

## Optimization Tips

1. Use incremental loading to reduce data volume
2. Enable caching for frequently accessed reference data
3. Partition large datasets by date or key
4. Use columnar formats (Parquet) for intermediate storage
5. Monitor skew and repartition as needed

## See Also

- [Runtime Scaling](/docs/runtime/scaling) — Auto-scaling configuration
- [Caching](/docs/runtime/caching) — Caching details