Apache Cassandra Developer Associate
More study resources
Cheat sheet
Keep the key ideas within reach.
ExploreKey terms
Get comfortable with the vocabulary.
ExploreStudy plan
Give each week a clear focus.
ExploreCommon mistakes
Understand the traps and how to avoid them.
ExploreCase studies
Put your knowledge to work in context.
ExploreExam-day guide
Prepare for the day ahead.
ExploreAbout this exam
The Cassandra Developer Associate credential is an entry-level certification for software engineers who build applications on Apache Cassandra rather than administer clusters. It signals that you can model data around query patterns, write correct CQL, and use a client driver without fighting the database's distributed nature. Candidates typically have several months of hands-on experience writing Cassandra-backed services and understand partition keys, replication, and consistency trade-offs well enough to defend a schema choice in review.
The exam samples six areas: architecture and data distribution (ring, tokens, replication, gossip, hinted handoff), data modeling and schema design, CQL, client drivers and application development, performance tuning and query optimization, and operations and troubleshooting. Expect scenario questions rather than trivia: given an access pattern, which primary key avoids a hot partition; given a read latency spike, which metric or log to inspect first. Multiple choice is the format.
Preparation works best in that order: build a small three-node cluster locally, model a realistic domain (messaging, IoT readings, time-series events), then deliberately break things - wide partitions, ALLOW FILTERING, tombstone storms - and read the logs. The official Apache Cassandra documentation, the CQL reference, and driver manuals are the authoritative study sources. The certifying body publishes current exam logistics, so check its site for scheduling and policy details.
What you'll study
About Apache Cassandra Developer Associate preparation
What format is the exam?
It is delivered as multiple choice. Expect scenario-style items that describe an access pattern, a schema, or an error and ask you to pick the best action. There is no hands-on lab component, but questions assume you have actually run Cassandra and can reason about cluster behavior rather than recite definitions.
How difficult is it for someone new to Cassandra?
Moderate. Developers with relational experience often fail on data modeling first, because normalization instincts produce bad partition keys. If you have shipped one Cassandra-backed service and can explain why you chose each primary key, you are close. Budget several weeks of deliberate practice on a local cluster rather than passive reading.
How long is the exam and how many questions?
The certifying body sets the current question count and time limit, and they change between releases. Check the official certification page for the authoritative figures before you schedule. Do not plan pacing around numbers from older blog posts or study guides, which are frequently out of date.
What does it cost and what is the passing score?
Fees and the passing threshold are published by the certifying organization, not fixed by the exam's content. Consult its official site for current pricing, any regional variation, and the score report format. Third-party claims about exact cut scores are unreliable, so treat the official page as the only source.
If I fail, can I retake it?
Retake rules, waiting periods, and any additional fees are set by the certifying body and are not part of the technical content. Check its official policy page before booking. Practically, candidates who fail usually misjudge data modeling or consistency questions, so review those domains before scheduling a second attempt.
How long is the certification valid?
Validity periods and any recertification requirements are defined by the certifying organization. Confirm the current policy on its official site, since programs revise these terms. Because Cassandra evolves - vector search, accord, storage-attached indexes - assume you will need to refresh knowledge regardless of formal expiry.
What should I study first?
Start with architecture: the ring, tokens, replication factor, and consistency levels. Then data modeling, because every later domain depends on partition and clustering keys. CQL comes next, then drivers, then tuning and operations. Studying CQL syntax before understanding partitioning leads to memorization without the judgment the scenario questions reward.
What are the most common failure modes?
Choosing partition keys that create unbounded or hot partitions; using ALLOW FILTERING as a fix; misunderstanding quorum math and how it interacts with replication factor; confusing compaction strategies; and misreading tombstone warnings. Another frequent trap is assuming a secondary index behaves like a relational index on a high-cardinality column.
Do I need to know cluster administration?
Not deeply, but operations questions appear. You should recognize what hinted handoff, repair, gossip, and compaction do, and how to read nodetool output and logs when diagnosing latency or unavailable errors. You do not need to design a multi-datacenter deployment, but you should know how replication across datacenters affects consistency choices.
Which resources are authoritative?
The Apache Cassandra project documentation, the CQL reference, and the official DataStax drivers' manuals are the primary sources. Use them alongside a local cluster so you can verify claims. Avoid question dumps: they are often outdated or wrong, and memorizing them builds none of the reasoning the exam tests.