We have deployed cassandra-cluster in EKS. For some days we are getting alert messages

Pod cassandra-0 Memory usage is 90.08 percentage in last 2 minute

We checked cassandra pod logs, we were not able to find any error at that time.

INFO 01:25:44 Redistributing index summaries

INFO 01:50:48 3 prepared statements discarded in the last minute because cache limit reached (6062080 bytes)

Is there any way we could reduce the memory usage for this process or can we run this process in a cronjob ?


1 Answer 1


"Redistributing index summaries" is part of the normal operation of of the cluster and is performed by an IndexSummaryManager thread in Cassandra, not a separate process.

Cassandra maintains summaries of SSTable indices which get resized periodically in order to allocate more memory to "hot" SSTables. The memory pool for index summaries is a fixed size from the heap that is either (1) set as index_summary_capacity in cassandra.yaml, or (2) 5% of the heap size.

You can reduce the size of the index summary memory pool BUT since it is allocated from heap, (3) it would be pointless to do so, AND (4) would end up hurting the performance of your cluster.

You should instead review the resources allocated to your pods. There are no hard rules but we generally recommend provisioning at least 32GB of RAM so you could configure the heap size to 16-24GB for production workloads. Cheers!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.