Offline Partitions Indicate Total Partition Unavailability
criticalWhen partitions go offline (all replicas unavailable), they cannot serve produce or consume requests, causing complete unavailability for affected data.
Monitor kafka.replication.offline_partitions_count > 0 or kafka.partition.offline = 1 for any partition. This indicates complete partition failure requiring immediate attention.
1. Identify affected topics: Determine which topics have offline partitions. 2. Check all replicas: Verify status of all replicas for offline partitions. 3. Restore replicas: Bring broker(s) hosting replicas back online. 4. Check for data loss: If unclean.leader.election.enable is false, may need to enable temporarily to restore availability. 5. Review logs: Check broker logs for errors causing replica failures. 6. Verify disk health: Ensure disk/storage is functional on affected brokers. 7. Consider increasing replication factor: Add more replicas to prevent future offline scenarios.