A bank is using Amazon Managed Streaming for Apache Kafka (Amazon MSK) to populate real- time data into a data lake. The data lake is built on Amazon S3, and data must be accessible from the data…

Tripling the broker storage is a reactive solution for an imminent disk space issue, but the question asks what to do to *prevent* running out of space, which implies a proactive measure. Additionally, MSK storage scaling typically does not require a cluster restart.

While an immediate storage increase might be needed, to *prevent* the cluster from *running out* of disk space in the future, the specialist should create or refine a CloudWatch alarm on the `KafkaDataLogsDiskUsed` metric. This allows for proactive notification at an earlier threshold (e.g., 80% instead of 90%), enabling timely scaling of broker storage or adjustment of retention policies before capacity becomes critical.

Creating a custom Amazon MSK configuration can adjust Kafka parameters like log retention, but it is not the primary mechanism for scaling underlying EBS storage capacity to handle increased data volume, nor is it the most direct 'preventative' measure against full disks compared to monitoring.

Tripling the number of consumers helps process data faster but does not reduce the amount of disk space Amazon MSK uses to retain data according to its configured retention period; Kafka stores data regardless of consumer speed for its log retention settings.