[BUG][opensearch] java heap size always 1GB

[tmoody1]

Describe the bug
The java opts flag doesn't seem to change the heap size in elasticsearch. The only way we've found to change this is to overwrite the config file manually. Not sure if this is by design or not, if it deliberately set to 1GB please close this.
To Reproduce
Steps to reproduce the behavior:

1. set esJavaOpts to 2GB
   esJavaOpts: "-Xms2048m -Xmx2048m"
2. Check the heapsize on the installed chart
   GET _cat/nodes?h=heap*&v

Expected behavior
I expect the heapsize to be 2GB for the nodes that have esJavaOpts set to 2GB (in our case the data nodes)
what actually happens is all nodes show 1GB

Chart Name
opensearch

what seems to happen is instead of taking the JAVA_OPTS environment variable elasticsearch instead uses the value set in the file /usr/share/opensearch/config/jvm.options which has

-Xms1g
-Xmx1g

Host/Environment (please complete the following information):

• Helm Version: 3.6.3
• Kubernetes Version: 1.22.1

Additional context
Prob not really relevant to the bug as a whole but...
Issue was spotted our side when we had a particularly large index, this triggered the circuit breaker exception in ES.
looking online we saw this could be solved by simply increasing the heapsize in ES which led to this bug. We have since reduced the size of our indicies which seems to have prevented the circuit breaker exception.

[peterzhuamazon]

This is more of a opensearch-build issue than helm charts. Transfer issues.

[peterzhuamazon]

Talked to the team it is more of a core issue, transfer to OpenSearch.

[tmoody1]

I think you can close this actually.
Later versions of the helm chart seem to set limits correctly