Confluent Certified Administrator - Study Strategy¶

Study Approach¶

Phase 1: Kafka Architecture Deep Dive (1-2 weeks)¶

1. Core Architecture
2. Understand broker internals (log segments, indexes, compaction)
3. Master replication mechanics (ISR, leader election, failover)
4. Learn ZooKeeper role and KRaft migration path

5. Study controller responsibilities and metadata management

6. Hands-on Setup

7. Deploy a multi-broker Kafka cluster (3+ brokers)
8. Configure ZooKeeper ensemble or KRaft quorum
9. Practice broker configuration and tuning
10. Explore log directories and segment files

Phase 2: Operations and Management (2-3 weeks)¶

1. Cluster Operations
2. Practice topic creation, modification, and deletion
3. Execute partition reassignment and preferred leader election
4. Perform rolling upgrades in a test environment

5. Practice broker decommissioning and addition

6. Configuration Management

7. Study all broker-level configurations
8. Understand topic-level configuration overrides
9. Learn client quota management
10. Practice dynamic configuration changes

Phase 3: Monitoring, Security, and Platform (2-3 weeks)¶

1. Monitoring and Troubleshooting
2. Set up JMX monitoring with Prometheus/Grafana or similar
3. Practice interpreting key metrics
4. Simulate failure scenarios and troubleshoot

5. Monitor consumer lag and identify bottlenecks

6. Security Configuration

7. Configure SSL/TLS for encryption
8. Set up SASL authentication (SCRAM recommended for practice)
9. Implement ACLs for authorization

10. Practice end-to-end secure cluster deployment

11. Confluent Platform

12. Deploy Control Center and explore monitoring features
13. Configure Schema Registry for high availability
14. Manage Connect clusters via REST API
15. Deploy and manage ksqlDB servers

Phase 4: Exam Preparation (1 week)¶

1. Practice Exams
2. Take practice exams and review incorrect answers
3. Focus on operations and configuration questions

4. Review domain weightings and adjust study focus

5. Final Review

6. Review all critical metrics and their alert thresholds
7. Review CLI commands for topic and ACL management
8. Review security protocol combinations

Study Resources¶

Official Resources¶

• 📖 Kafka Operations - Operations reference
• 📖 Confluent Admin Guide - Administration guide
• 📖 Confluent Training - Official training courses
• 📖 Kafka Configuration - All configuration reference

Recommended Courses¶

• 📖 Administration with Confluent (Confluent) - Official admin course
• 📖 Kafka Internals - Free foundational courses
• 📖 Confluent Operations Training - Hands-on operations

Practice and Hands-on¶

• 📖 Docker Compose Setup - Quick local cluster
• 📖 Kafka CLI Tools - Command-line tools reference
• 📖 Confluent CLI - Confluent CLI reference

Exam Tactics¶

Question Strategy¶

1. Configuration Questions - Know defaults and when/why to change them
2. CLI Command Questions - Know exact syntax for common operations
3. Troubleshooting Questions - Map symptoms to root causes
4. Architecture Questions - Understand component interactions

Time Management¶

• 1.5 minutes per question average
• Flag and move - Do not spend more than 2 minutes on any question
• Reserve 10-15 minutes for reviewing flagged questions
• Operations domain is 30% - Expect 18 questions on managing/operating

Domain Priorities (by weight)¶

1. Managing and Operating (30%) - Highest weight, focus on CLI and operations
2. Monitoring and Troubleshooting (20%) - JMX metrics and problem diagnosis
3. Confluent Platform (20%) - Control Center, Schema Registry, Connect
4. Kafka Fundamentals (15%) - Architecture and replication mechanics
5. Security (15%) - Authentication, authorization, encryption

Key Areas to Master¶

• Broker configuration defaults and tuning parameters
• CLI commands for topic management, ACLs, and consumer groups
• JMX metrics and their significance
• Rolling upgrade procedures
• Security protocol and SASL mechanism combinations
• Confluent Platform component architecture

Common Pitfalls¶

1. Confusing min.insync.replicas scope - Set at topic or broker level, not per producer
2. Partition count cannot be decreased - Only increased after creation
3. Rolling upgrade order matters - Update config, restart one by one, then update protocol
4. unclean.leader.election.enable - False by default, enabling risks data loss
5. ACL wildcard matching - * matches all resources of a type, not a regex
6. Security protocols - SASL_SSL is SASL auth + TLS, not two separate configs
7. Consumer lag - High lag alone is not always a problem, trend matters
8. Log compaction - Only applies to closed segments, not the active segment