My project have setup an Stateless Apache Ignite cluster in Kubernetes. During performance testing, the memory for each Apache Ignite pods has reach its memory limit and slows down the application performance.

After doing some reading, we discover that there is no any eviction policy implemented by default and there is actually a section from documentation that suggesting that we can implement a eviction for off-heap eviction. https://apacheignite.readme.io/docs/evictions#section-random-lru

However, we face a problem on implementing the policy because we are using helm chart and it does not provide much option for us to modify the configuration xml. https://github.com/helm/charts/tree/master/stable/ignite

Help would very much appreciated