Add resource requests/limits for elasticsearch, nifi, modernization-api

[EmmanuelNwa247]

Summary

NBS7 Helm charts for elasticsearch, nifi, and modernization-api either had only resources.limits (no requests) or oversized defaults for Small STLT clusters. When only limits are defined, Kubernetes auto-sets requests equal to limits — meaning Elasticsearch alone was guaranteeing 6GB on a node that only has 8GB. This caused node crashes on dev73 when multiple pods landed on the same node.

Ticket reference https://cdc-nbs.atlassian.net/browse/DEV-548

Before:

• elasticsearch: limits.memory 6Gi only → 6Gi guaranteed
• nifi: limits.memory 6Gi only (Octopus overrides to 1.5Gi on Skylight envs, but STLTs get the 6Gi default)
• modernization-api: limits.memory 4Gi only → 4Gi guaranteed
• Total guaranteed by just these 3: 16GB out of 19.5GB allocatable — leaving 3.5GB for every other service

Changes

Added resources.requests and right-sized resources.limits for Small STLT baseline (3 × m5.large):

Service	Request Mem	Request CPU	Limit Mem	Limit CPU
elasticsearch	1536Mi	500m	3Gi	1000m
nifi	1536Mi	500m	3Gi	1000m
modernization-api	768Mi	250m	2Gi	500m

Also reduced nifi jvmheap from init=4g/max=4g to init=1g/max=2g to stay within the new container limit.

After: Total guaranteed by these 3: ~3.75GB — down from 16GB.

Testing