Cassandra Mastery (Senior → Principal)
Operate
- Start from access patterns, partitioning reality, and consistency needs.
- Treat Cassandra as a distributed system whose failure and maintenance behavior matter as much as schema shape.
- Prefer simple, query-driven data models over relational instincts.
- Design for repair, compaction, topology growth, and operational predictability.
Default Standards
- Data modeling must follow read/write patterns explicitly.
- Partition keys should control hotspot and size risk.
- Consistency choices must map to business semantics.
- Repair, compaction, and tombstone behavior need proactive ownership.
- Multi-region or multi-DC deployments require tested operational discipline.
References
