Apache Cassandra Data Modelling and Operations Training Course

Course overview

Apache Cassandra is often selected for high availability and multi-region scale, then undermined by relational data models, oversized partitions, inefficient secondary indexes, or inconsistent operational practices. Engineers need to translate application access patterns into partition keys, clustering columns and denormalised tables that remain predictable under load. Operations teams must also understand how replication, consistency levels, compaction, repair and tombstones interact before they become latency incidents, storage growth or unavailable data.

This five-day Apache Cassandra Data Modelling and Operations Training Course teaches participants to design query-led CQL schemas, estimate partition sizes, select replication strategies and apply consistency levels to real service requirements. Participants work with Cassandra’s write path, read path, SSTables, memtables, commit logs, compaction strategies and anti-entropy repair. They use cqlsh and nodetool to inspect cluster state, diagnose common failure modes, plan capacity, manage schema changes, tune operational settings and establish monitoring signals that support reliable production decisions.

Instructor-led explanations are paired with guided labs on a working Cassandra cluster, design reviews and incident-based exercises. Participants model a multi-region application workload, test query behaviour, investigate simulated performance and availability issues, and develop an operational runbook. They leave with a documented Cassandra data model, access-pattern matrix, replication and consistency rationale, capacity assumptions, monitoring checklist and repair/maintenance plan that can be adapted for their own platform. The course is delivered in the classroom or live online over one week and includes a certificate on completion.

Course objectives

By the end of this course, participants will be able to:

• Model Cassandra tables from application query patterns using partition keys, clustering columns and deliberate denormalisation
• Calculate partition-size and cardinality estimates to identify hot partitions and unbounded data-growth risks
• Write and test CQL queries, prepared statements and schema definitions for predictable read and write paths
• Configure replication factors and consistency levels for single-region and multi-datacentre availability requirements
• Diagnose cluster health with nodetool outputs, system tables and Cassandra log evidence
• Select compaction, compression, TTL and tombstone-management settings for defined workload characteristics
• Plan repair, backup, restore and rolling-maintenance procedures that protect data availability
• Produce a Cassandra operational runbook containing monitoring thresholds, capacity assumptions and incident actions

Benefits of attending

Target competencies

Who should attend

• Database Administrators — who operate Cassandra clusters and need repeatable maintenance, repair and recovery procedures
• Data Engineers — who design high-volume data stores and must make query-led schema decisions
• Backend Software Engineers — who build services on Cassandra and need to avoid inefficient partitions and queries
• Site Reliability Engineers — who monitor distributed databases and respond to latency, disk and node-failure incidents
• Platform Engineers — who provide managed Cassandra services and need sound topology, scaling and observability practices
• Technical Architects — who assess Cassandra fit, model trade-offs and define multi-region data availability patterns

Requirements and prerequisites

Participants should have practical experience writing basic SQL or CQL-style queries and understand tables, primary keys, indexes and application data flows. Familiarity with Linux command-line navigation, reading log files and using a terminal is expected, as labs use cqlsh and nodetool. Prior exposure to distributed systems concepts such as replication, latency and eventual consistency is helpful, but the course explains Cassandra-specific behaviour from first principles. Participants do not need prior Cassandra administration experience, Java development skills, Kubernetes expertise or a production cluster of their own.

Training methodology

The course combines short instructor-led technical briefings with structured labs on a Cassandra cluster. Participants use cqlsh to create and query schemas, then use nodetool and cluster evidence to investigate simulated imbalance, tombstone and node-availability scenarios. Small-group design reviews compare alternative models against stated access patterns, latency targets and regional resilience needs. Each day closes with an applied task, culminating in an end-of-course implementation plan that connects data model decisions, operating procedures, monitoring and capacity assumptions to a participant’s own workload.

Course outline

Day 1: Cassandra architecture and query-led modelling

• Cassandra ring architecture, tokens and virtual nodes
• Coordinator nodes and the distributed request path
• CQL primary keys: partition keys and clustering columns
• Query-first modelling and access-pattern inventories
• Denormalisation patterns for Cassandra tables
• Partition cardinality, size estimates and hot-partition risks
• Schema creation and inspection with cqlsh

Workshop: Model an order-tracking workload from an access-pattern matrix and produce CQL table definitions with documented key choices.

Day 2: Replication, consistency and resilient reads and writes

• Replication strategies: SimpleStrategy and NetworkTopologyStrategy
• Replication factor selection by datacentre
• Consistency levels for reads, writes and serial operations
• QUORUM, LOCAL_QUORUM and LOCAL_ONE trade-offs
• Read repair, speculative retry and digest reads
• Lightweight transactions and compare-and-set use cases
• Multi-datacentre topology and failure-domain design

Workshop: Configure and test consistency-level scenarios for a multi-datacentre service, then produce an availability and data-consistency decision record.

Day 3: Storage engine behaviour and performance tuning

• Write path: commit log, memtable and SSTable creation
• Read path, bloom filters, indexes and partition lookups
• Size-tiered, leveled and time-window compaction strategies
• Tombstones, TTLs and deletion-marker failure patterns
• Compression settings and SSTable storage trade-offs
• Secondary indexes, storage-attached indexes and materialized-view considerations
• Prepared statements, paging and application query behaviour

Workshop: Investigate a slow-query case caused by tombstones and poor partition design, then produce a revised schema and compaction recommendation.

Day 4: Cluster operations, maintenance and recovery

• Cluster inspection with nodetool status, describecluster and tablestats
• Node bootstrap, decommission and replacement procedures
• Anti-entropy repair concepts and repair scheduling
• Incremental repair, full repair and consistency implications
• Snapshot, backup and point-in-time recovery considerations
• Rolling restart and configuration-change planning
• Disk, heap, garbage collection and thread-pool operational signals

Workshop: Work through a node-loss and disk-pressure incident, producing a sequenced recovery, repair and validation runbook.

Day 5: Monitoring, capacity planning and production readiness

• Cassandra metrics for latency, errors, dropped messages and pending tasks
• Prometheus metric collection and alert design
• Log-based diagnosis of timeouts, unavailable errors and gossip issues
• Capacity planning for data volume, replication and compaction overhead
• Load testing and baseline-performance measurement
• Security controls: authentication, authorisation and client encryption
• Production-readiness reviews and operational ownership

Workshop: Create a production readiness pack containing a monitoring dashboard specification, capacity forecast, alert thresholds and 90-day operating plan.

Tools & standards covered

Apache Cassandra, cqlsh, nodetool, Prometheus

A typical training day

Live online deliveries follow the same structure in the East Africa Time zone, with shorter screen blocks and longer breaks.

What the fee includes

How you can take this course

Certification

Participants who complete the full five days receive the Skillset Development Certificate of Completion, stating the course title, course code, dates and delivery format — suitable for professional-development records and employer reimbursement.

Frequently asked questions