Confluent CCAAK Glossary: Kafka Cluster Terms
April 13, 2026
Confluent CCAAK glossary of Kafka cluster, durability, config, security, and monitoring terms.
Use this glossary when Kafka administration, durability, and monitoring terms start to blur together. Keep it beside the cheat sheet and resources instead of using it as a substitute for troubleshooting practice.
| Term |
Short meaning |
| ISR |
In-sync replicas currently caught up enough to participate in safe acknowledged writes |
| Replication factor |
Number of copies of a partition spread across brokers |
| Leader election |
Process of choosing the active writable replica for a partition |
| Controller |
Kafka broker role that manages cluster metadata and partition state changes |
| Follower |
Replica that copies data from the leader and may join ISR when caught up |
| advertised.listeners |
Broker endpoints that clients are told to use when connecting back to the cluster |
| listeners |
Local broker bind addresses and protocols used for incoming connections |
| Under-replicated partition |
Partition whose replicas are not fully caught up to the leader |
| Min in-sync replicas |
Minimum number of caught-up replicas required for certain durable writes |
| Segment |
Kafka log file chunk stored on disk |
| Retention policy |
Rule controlling how long or how much log data is retained |
| Throttle |
Rate-limiting control applied to replication or client traffic |
| ACL |
Access control list used to allow or deny Kafka operations |
| Consumer lag |
Difference between the latest produced offset and a group’s committed position |
| JMX |
Java Management Extensions interface commonly used for Kafka metrics collection |
| Unclean leader election |
Availability-biased recovery behavior that can increase data-loss risk |
| Control Center |
Confluent management and observability surface rather than Kafka core itself |
| Schema Registry |
Confluent schema-management component for serialization contracts and compatibility |
| Kafka Connect |
Integration runtime for source and sink connectors |
Commonly confused pairs
| Pair |
Keep this distinction clear |
| partition vs replica |
unit of parallel data flow versus extra copy for resilience |
| leader vs controller |
active writable replica for one partition versus cluster metadata leader |
| follower vs consumer |
broker-side replica role versus client-side read process |
| replication factor vs min in-sync replicas |
total copies versus required caught-up copies for durable acknowledgement |
| broker config vs topic config |
cluster-wide default versus per-topic override |
| leader election vs rebalance |
partition leadership change versus consumer-group assignment redistribution |
| listeners vs advertised.listeners |
bind location versus client-facing endpoint |
| retention vs compaction |
timed or sized cleanup versus latest-key preservation |
| lag vs throughput |
backlog visibility versus processing rate |
| ACL vs SASL/TLS |
authorization versus authentication/encryption layers |
| Kafka core vs Connect/Schema Registry |
broker-admin behavior versus surrounding Confluent platform components |
Fast boundary reminders
| If the term really points to… |
Think of it as… |
| durability and safe writes |
ISR, replication factor, min ISR, unclean leader election |
| broker and cluster roles |
controller, leader, follower |
| connectivity path |
listeners, advertised.listeners, TLS, SASL |
| observability and diagnosis |
URP, offline partitions, lag, JMX |
| adjacent platform components |
Connect, Schema Registry, Control Center |
If the confusion is really about…
Revised on Sunday, May 10, 2026